RCT #177 - Do the highest bang-for-buck work. Plus: What to say after a failure; GitHub Copilot’s productivity benefits; Research data services landscape at Canadian and US institutions; What modern NVMes can do.
Do the highest bang-for-buck work. Plus: What to say after a failure; GitHub Copilot’s productivity benefits; Research data services landscape at Canadian and US institutions; What modern NVMes can do.
The Importance Of A High Research Productivity Focus
After a longer gap than I intended, I want to keep talking about this diagram, the flywheel which defines our teams’ missions.
Last issue I talked about nurturing our best clients first as a way to add work that we know is more likely to have significant research impact for a reasonable investment in our team’s time and resources.
And in fact, this idea that we should aim for the most research impact possible given our resources and constraints, a fundamental guiding principle. It has to be the beacon we steer by when making decisions.
We can use different words or phrases for this - bang-for-buck; cost-benefit; I have a mischievous habit of sometimes using the term “profitability”, e.g. the results minus the inputs, just to make people squirm.
But “Research Productivity” is probably the best term I’ve seen, for two reasons:
- It has a very clear definition in principle, even if it’s devilishly hard to measure: research impact divided by funding and other resources applied.
- It’s the same kind of wording used by funders and institutions when assessing researchers and their labs. We’re part of that world.
High Research Productivity is Our Duty To Science, Our Institution, and Our Teams
Research is scandalously underfunded. Our teams are entrusted with some of that funding. It’s our job to steward that all-too-scarce technical research support funding and resources placed under our control.
It’s our duty to science to make sure that those resources are applied to where it can do the most good, even if that sometimes means uncomfortable conversations and decisions. “But we’ve always done it this way” is not a good enough reason to put resources somewhere unlikely to pay off. Avoiding uncomfortable discussions with long-term clients isn’t a valid excuse for misallocating some of this scarce research support money.
It’s also our duty to our institutions. Most of us are in teams whose job it is to support the research aspirations of a particular institution, a particular community. For the same reasons as above we owe it to our institutions to apply our team’s efforts and our resources where it has the greatest benefit.
Finally, it’s our duty to our teams, for two reasons. One is professional pride. Our highly-qualified, over-worked, under-paid team members stayed in the world of research for a reason. They wanted to have a job that had real research impact. They have the skills and the drive to make that impact. They deserve to see their efforts have the impact they’re capable of.
And secondly:
High Research Productivity Is How We Justify Getting Resources
Our team members also deserve for their jobs to be secure, and to have more colleagues doing more things so they can be a little less overworked. They deserve to be able to grow and learn from each other, and to take on bigger challenges.
The research impact our team has is what justifies its funding.
Demonstrating the highest research productivity, the highest bang-for-buck, is the best possible argument for allocating more of those too-scarce technical research support funds to our team. It’s how we can grow our teams, and give our team members more professional growth opportunities.
The ruler we’re measured against is not “did you accomplish something with those resources”. It’s what those resources could have accomplished deployed elsewhere. The measuring stick, the unit of comparison, is the very best best research teams #173.
It may be uncomfortable to think of ourselves as vendors #123, but we are. We provide services and access to resources in exchange for funding.
In particular, we’re Professional Services firms, not Utilities (#127).
Utilities just mindlessly mass produce and provide undifferentiated power or water to whoever turns on a switch or opens a valve, and is willing to pay for the service. Utilities are cost centres, and are replicable if an alternative with lower costs come around.
But we’re professional providers of services. We bring expertise and judgement to bear, to help researchers accomplish as much impact for Science, Research and Scholarship and for our institutions. We make sure our team members can do their best work, and that the researchers can do theirs. That means applying judgement to what we work on.
Sometimes This Means Saying No or Phasing Out Work
And sometimes that means saying No to projects or kinds of work, even if would be useful to do them, even if we used to do them before.
I’ve been encouraging us to say No for some time (#56). Saying No is the essence of strategy, of having a focus. Every “yes” is an implicit “no” to all the other things that could have been done during that time or with that energy; it’s best to be explicit about our “no”s. I tried to provide some scripts in #131.
We Can And Must Hold Ourselves to High Standards
Even when we’ve made good choices about what kinds of work to do, it’s up to us, as much as our VPRs or anyone else, to hold ourselves to high standards (#165) for actually doing it well if we’re going to have high research productivity.
We’re the ones who really understand what our team’s potential is, what kinds of impact it could have. For most of our teams, if we’re not holding ourselves up to high standards of impact, no boss will come to tell us to do anything different — but our modest impact will be noted, and any requests for additional funding will be weighed accordingly.
Technical research support standards are rising #161. To keep up we need to make sure we’re learning what works and what doesn't, incorporating that into modest amounts of process and automation (#95), which increases our leverage and helps us get reproducibly strong results (#157) while letting our team members focus on the creative parts of the work.
Uncertainty And Long Time To Payoff Is Not An Excuse
None of this means we need to think only in the short term; nor am I living in a fantasy world where you turn some crank and then deterministically out pops some high-impact paper.
Research is a long game. Helping some junior new faculty member (or a faculty member new to doing computing-power researched) get over the initial barrier to entry to doing a new kind of work might not have any impact this year or next, but could make a huge difference for over the course of five years, for a comparatively modest short-term investment of effort.
And teaching “how to use the cluster” or “use R to analyze data” or software carpentry classes to grad students certainly doesn’t have any immediate research impact, but it can greatly cut down the costs to support trainees over the course of their work.
Yes, the payoff is over the long term, and it’s uncertain. You don’t know if this grad student or this postdoc is going to be the one to have an amazingly successful project. Heck, they don’t know that, either.
But uncertainty and time horizons don’t absolve us from the responsibility of stewardship over the resources entrusted to us. They doesn’t excuse us from having to think judiciously and carefully about how we and our team spends their time, how we marshal our efforts.
Whether in training (#162) or more broadly (#163), it’s up to us to measure and keep an eye on what matters - research impact - and have a plan, a theory of change, as to how our efforts drives that research impact. And to continually update that plan as we learn new things.
The fact that research is hard, uncertain, and slow is more reason, not less, to be thoughtful and careful about planning what work to take on so that we can do the most good.
And you know what? If we’re not sure? Researchers can tell us what has the highest research impact (#158). We just need to talk to them.
Finally, on that note of researchers are the best judge of research impact per unit dollars - Next issue, I’m going to share in some detail my absolutely most cancel-able technical research support opinion. That is: if researchers aren’t even in principle willing to pay what it costs for us to offer a service, we probably shouldn’t be doing it.
With that, on to the roundup!
Managing Teams
In the previous issue of Manager, Ph.D. I talked more about the advantages and challenges STEM PhDs have in becoming capable leaders, looking in particular at a paper on (industrial) R&D leadership. Again, we see that there are real, advanced, strengths compared to non-R&D leaders, but weaknesses in some of the basics. But the basics are things we can work on.
The roundup included articles on:
- How imposter syndrome affects individual contributor productivity
- How to build trust (I really like this one)
- Managing the ups and downs of team performance
- Leadership myths
- The struggles we face going from a star performer to a star [or even competent: JD] manager.
Technical Leadership
What to Say Next After a Team Setback: Beyond the ‘F’ Word - Karin Hurt and David Dye, Lets Grow Leaders
I like articles by these two for their list of helpful phrases and scripts. Sometimes it’s a lot easier to do the right thing if you have a pocket phrasebook of things to say while doing it.
I’ll offer a categorization for the phrases they suggest:
- Focus on learning
- “Let’s consider this a learning opportunity.”
- “What can can we take away from this experience?”
- Focus on next steps
- “This didn’t go as planned. What’s our next move?”
- “We Missed Our Objectives. How Can We Pivot?”
- Focus on we’re-all-in-this-together:
- “What can I do to be helpful here?”
- Focus on process:
- “What data or insights led us astray?”
I think their focus-on-process question is particularly important.
Over in Manager, PhD (#158, #165), but also here (#106, #153) I talk a lot about decision making under uncertainty, which is always the case for us in research (or hiring, or…).
You can’t use the outcome of a decision made under uncertainty to decide whether the decision itself was bad or good. The best decision available under the circumstances will sometimes lead to bad results, or you could luck out with a terrible decision that works well, because there’s so much you don’t know.
What you can do is look at the process that lead to the decision and see if there’s something to be learned there. Was there a relevant source of information that was missed? Was there an assumption that turned out untrue? If so, that’s something that can be used to inform future decisions.
But sometimes you and the team did everything right, and there just were unknown unknowns that tripped you up. It’s no fun at all, but it happens. All you can do in that case is the other stuff — learn, focus on next steps, and make it clear you’re all in it together.
And honestly, a shared fiasco that you and the team handle well together can really improve the team’s mutual trust and esprit de corps. One that’s handled poorly can shred both of those things.
Research Software Development
Measuring GitHub Copilot’s Impact on Productivity - Ziegler et al, Communications of the ACM
[disclaimer - my day job right now is working for a company best known currently for making AI hardware, so take this all with a grain of salt. For what it’s worth, though, I started playing with Copilot before joining my current employer, and enjoying it a great deal.]
Whether it’s Copilot per se, or locally hosted models like StarCoder 2 or Code Llama, it’s becoming increasingly clear that these tools are very productivity-enhancing for a lot of kinds of software development, once you get used to using them. Heck, their own internal tool noticeably improved Google developers’ productivity, models have only improved by then (July 2022!), and whatever one might say about Google as a company they are starting from a baseline of extremely capable developers.
A lot of people poke around at tools like Copilot and leave disappointed because their initial impressions are of code suggestions with clear weaknesses. The value of these tools, though, isn’t that it blatts out reams of perfect code for you. From one of the Key Insights listed in the paper:
While suggestion correctness is important, the driving factor for these improvements appears to be not correctness as such, but whether the suggestions are useful as a starting point for further development.
That’s been my experience - in my own personal projects, these tools let me me see and assess starting points much more quickly, and even rapidly mock out a few approaches first to see how they’d look before settling on one.
The other thing is:
The reported benefits of receiving AI suggestions while coding span the full range of typically investigated aspects of productivity, such as task time, product quality, cognitive load, enjoyment [my emphasis added], and learning.
I find these tools just fun, taking coding from being a purely solitary thing and turning it into a conversation with an incredibly eager to please junior colleague who is naive and thinks about things “weirdly” but also occasionally offers surprising insight.
While the authors looked at a number of metrics, the most obvious and seemingly simplistic one, “completion acceptance rate”, seemed to be the most reliable measure for productivity gains, and these numbers seemed to average at around 21%-23%. So even the developers which reported that these tools made them feel subjectively more productive rejected the suggested autocompletions 4 times out of every 5.
Research Data Management and Analysis
The Research Data Services Landscape at US and Canadian Higher Education Institutions - Ruby MacDougall, Dylan Ruediger, ITHAKA S&R
An interesting overview of digital research support services at 120 US and 8 Canadian institutions, through the lens specifically of data services.
One of the things that struck me about this work was the sheer number of services being offered, especially at R1 schools:
That’s north of 30 different services some institutions have, and if you look at the breakdown by providers, it’s about 45% by the libraries, 25% by the research office (I assume that’s digital research support teams under the VPR, but I’m not sure), and then medical schools, IT, and individual academic departments making up the rest.
This explains the authors’ call for greater coordination and collaboration at R1s, and broader support at R2s/liberal arts college, and I agree. Consolidation isn’t the right answer here — no one’s interest is best served by having the management of actively updated (say) single-cell transcriptomics datasets being taken over by the libraries — but with consultation and training making up almost all of the services, there’s a lot of value in making sure that we’re making things easy for researchers while making the best use of every teams’ particular skill sets.
(I also think this is part of a long-term trend as digital-powered research gets both more diverse and more ubiquitous, of increasingly numerous, focussed, and specialized smaller teams cropping up, rather than the one-huge-team-that-does-everything. It doesn’t surprise me that we’re seeing this first and most strongly in data, since as a friend of mine often says, “there’s no such [single] thing as data”. Of course, there’s also no such single thing as research software or research computing, although the community seems to be a little further behind on those realizations).
Finally, it sort of pains me to see how “stats help desk” type services are languishing in this big data age, when statisticians obviously have a lot of useful perspective to bring to a lot of this work.
Research Computing Systems
What Modern NVMe Storage Can Do, And How To Exploit It: High-Performance I/O for High-Performance Storage Engines (PDF) - Gabriel Haas and Viktor Leis, VLDB 2023
Interesting and in-depth paper covering how nontrivial it is to get peak performance is out of NVMe storage. There’s way too much to summarize here, and the discussion is (naturally) focussed on database use rather than (say) local staged-in storage for compute jobs, but I think their conclusions have a lot of useful tidbits for us, including:
- Arrays of NVMe SSDs can achieve the performance promised in hardware specs
- The best trade-off between random IOPS, throughput, and I/O amplification is achieved with 4KB pages
- I/O should be performed directly by worker threads
- Linux kernel I/O interfaces performed surprisingly well for IOPS, while io_uring was the only way to get full bandwidth.
Random
This is this newsletter’s second leap year. And while schadenfreude is certainly wrong and beneath us, nonetheless let’s read together a round-up of leap-day bugs.
Google groups has ended Usenet support.
Exploring the slightly nebulous meaning of git’s current branch.
Want to buy your own (emulated) 30 qubit quantum computer? Or a real 3-qubit desktop quantum computer?
Learn postgres in the browser.
The sad tale of Voyager 1 going mad and dying.
(Almost) frying a $500M Mars Rover.
An online electronic connector identifier. GPT in 500 lines of SQL? Sure, that’s cool I guess, but why not GPT-2 in Excel? Maybe that’s how you can apply AI to managing the 20,000 car parts your Formula 1 team needs to keep track of.
As you know, this is principally an embedded database fan club newsletter, so recent embedded DB news - Cloud Backed SQLite, (x2), an embedded ClickHouse OLAP engine, and using DuckDBs JSON support to replace jq. I used jq for something earlier this week, the syntax for anything nontrivial is still baffling to me, should have used this.
That’s it…
And that’s it for another week. If any of the above was interesting or helpful, feel free to share it wherever you think it’d be useful! And let me know what you thought, or if you have anything you’d like to share about the newsletter or stewarding and leading our teams. Just email me, or reply to this newsletter if you get it in your inbox.
Have a great weekend, and good luck in the coming week with your research computing team,
Jonathan
About This Newsletter
Research computing - the intertwined streams of software development, systems, data management and analysis - is much more than technology. It’s teams, it’s communities, it’s product management - it’s people. It’s also one of the most important ways we can be supporting science, scholarship, and R&D today.
So research computing teams are too important to research to be managed poorly. But no one teaches us how to be effective managers and leaders in academia. We have an advantage, though - working in research collaborations have taught us the advanced management skills, but not the basics.
This newsletter focusses on providing new and experienced research computing and data managers the tools they need to be good managers without the stress, and to help their teams achieve great results and grow their careers.
Jobs Leading Research Computing Teams
This week’s new-listing highlights are below in the email edition; the full listing of 104 jobs is, as ever, available on the job board.
Technical Lead (Human Genome Informatics) - Australian BioCommons - University of Melbourne, Melbourne VIC AU
The Technical Lead will drive the design and implementation of large-scale digital platforms to transform human genomics data management in Australia, aiming to advance human health and biomedicine. A major focus of this role will be collaboration with national and international medical research organisations and infrastructure providers.
Head of Bioinformatics Analytics - Thyme, London UK
We are a pioneering biotechnology company dedicated to advancing healthcare through cutting-edge research and innovation.
Our team is comprised of talented individuals who are passionate about harnessing the power of bioinformatics to drive impactful discoveries and improve patient outcomes. We are seeking a dynamic and experienced individual to join us as the Head of Bioinformatics Analytics. In this role, you will lead our bioinformatics analytics team, overseeing the strategic direction, development, and execution of bioinformatics initiatives to support our research and development efforts.
Digital Research Computing Manager - Hays (Recruiter), Reading UK
You will be joining an organisation which has been at the forefront of UK higher education for nearly a century. They have become innovators and pioneers over the years, pushing academic boundaries and leading social change. This role sits within the Digital Research Computing Team, which provides specialist expertise in the provision and use of IT, computation and data analysis to support research and academic teaching. The role is accountable for the strategic engagement, vision, development and delivery of Digital Technology Services (DTS) in support of research. To proactively manage and co-ordinate between research groups and DTS, supporting executive and senior management across all areas to ensure suitable digital research services and appropriate digital research support processes are in place, documented, and available to support research within the organisation.
Scientific Computing Support Desk Manager and Engineer - Diamond Light Source, Oxfordshire UK
As the Scientific Computing Support Desk Manager and Engineer within Scientific Computing team you will play a crucial role in the development and support of the platforms and services that underpin the operations (Controls) and Scientific outputs at Diamond. To fulfil this role you will be an enthusiastic and capable problem solver able to head up the support desk team that provides first and second line infrastructure support to staff and users of the facility. The support desk you will lead is backed by a diverse team of subject matter experts including high performance computing and storage, cloud, Linux, and Microsoft services. Responsibilities of the role will include:
Director, Strategic Data Analytics - Abbott, Alameda CA USA
Our location in Alameda, California has an opportunity for a Director of Strategic Data Analytics. This position will oversee the data strategy, insights, and storytelling across the customer journey and marketing touchpoints, initiatives, and enterprise objectives. This individual will be responsible for looking end-to-end across our data ecosystem, enabling new capabilities, & managing ongoing operations. This leader directs a team that helps define, build, and optimize the connected data portfolio, acting as a central management function that operates across multiple teams, including quality, marketing, finance, commercial teams, strategy, etc. This individual should have a strong ability to communicate and will be accountable for translating analytics and data into strategic recommendations to senior leaders. Help advance the Global Market Insights team’s mission to combine empathy and analytics in a way that provoke bold actions to best serve our customers.
Director, Data Science & AI, Omics Data Lead - Merck, Cambridge MA USA
Our Artificial Intelligence and Machine Learning (AI/ML) capabilities are vital catalysts for our mission to invent new medicines that save and enhance lives. The Data, AI, and Genome Sciences (DAGS) function at our organization adopts an AI/ML-first approach to enhance target and biomarker discovery by driving the understanding of complex disease mechanisms. As the Director of Data Science and AI, Data Lead, you will leverage your expertise in genomic technologies, multi-omics data, and computational and AI/ML techniques to drive the development of a scalable, FAIR data infrastructure and evidence-based knowledge graph. You will collaborate with Computational Biologists, Bioinformaticians, Data Scientists, Software Engineers, and AI/ML Engineers as part of a cross-functional team dedicated to identifying therapeutic targets. You will report to the Executive Director and Head of AI/ML.
Director, Quantum Computing Applied Research - NVIDIA, Santa Clara CA USA
At NVIDIA, we’re solving the world’s most exciting problems with our unique approach to accelerated computing. We’re looking for passionate technologists with deep quantum computing domain expertise to lead technical path-finding. As Director for Quantum Computing Applied Research, you'll lead our technical path-finding efforts, working with cross functional teams in Product, Engineering and Applied Research to develop innovative technologies that advance the state of Quantum Computing and intercept NVIDIA products. The Quantum Computing organization is a small, strong, and visible group both inside and outside of NVIDIA while Quantum Information Science is an exciting area to drive strategy. We need a self-starting leader to continue to grow this area. Do you have the rare blend of both technical and product skills with a passion for groundbreaking technology? If so, we would love to learn more about you!
Assistant Director of Research Computing - William & Mary, Williamsburg VA USA
Reporting to the Executive Director of Research Computing, the Assistant Director of Research Computing is responsible for assisting William & Mary (W&M) and Virginia Institute of Marine Science (VIMS) faculty/students with computing and computational science initiatives. This includes maintaining W&M/VIMS High Performance Computing (HPC) infrastructure as well as any new hardware/software used for Research Computing.
Executive Director for Research Initiatives, AI, Data, and Computing - Rice University, Houston TX USA
In collaboration with the Faculty Director of the Institute and its Advisory Board, the Executive Director develops plans to achieve the Institute’s mission and strategic priorities and is responsible for directing the implementation of these plans. The primary responsibilities of the Executive director are (a) fund-raising, which includes identifying, qualifying, and developing potential funding sources and opportunities that will bring additional research funding aligned with the Institute’s research directions, and developing and writing proposals to leverage these opportunities in collaboration with interested faculty members; and (b) community-building internally and externally which includes facilitating intersecting research communities within the Institute, connecting them with researchers at other institutions at the local, national and international level, and establishing long-term strategic relationships and partnerships for the Institute (with state and federal agencies, industry, foundations, philanthropic entities, universities, the Texas Medical Center, and other relevant organizations). In addition, the Executive Director, aided by the Director of Operations of the Institute, leads the alignment of resources with the Institute’s mission and strategy, oversees the Institute’s administrative workload and programmatic activities, staff hiring, as well as their needs, and a plan for staff growth.
Senior Manager, Data Exploration and Learning for Precision Health Intelligence - University of Utah, Salt Lake City UT USA
The Senior Manager will play a pivotal role in the strategic planning and execution of the DELPHI Initiative. This includes developing and implementing a wide range of strategic, proactive, and capacity building activities designed to enable and enhance data science research that includesstrategic research advancement, communication of research and research opportunities, enhancement of collaboration and team science, proposal development, and nurturing alliances between and among faculty at the U of U and funding agencies. A key function of this position is to work with stakeholders to develop program proposals, budgets (including a schedule for obtaining external support), goals, and implementation plans for review by health sciences center leadership. The individual in this position will be expected to operate independently to achieve program outcomes. In addition, this individual will contribute to emerging and ongoing SVPHS Research Unit initiatives and administrative responsibilities.
Lead Software Developer, Data Analytics for Canadian Climate Services - University of Toronto, Toronto ON CA
Under the direction of Prof Steve Easterbrook, and in close collaboration with our Lead Developer & Systems Architect, the incumbent is responsible for acting as a lead developer for a major scientific nationwide project funded by the Canadian Foundation for Innovation (CFI). The Data Analytics for Canadian Climate Services (DACCS) project is building a cloud-based platform to facilitate conversion of raw climate data from satellite observations and simulation models into relevant, credible and actionable information products for assessing climate impacts and risks. A key feature of DACCS is support for creating and sharing analytics workflows alongside the datasets they apply to. DACCS will integrate with existing international climate data services such as the Earth System Grid Federation, while providing additional services tailored to the Canadian context. To build the DACCS platform, the University of Toronto is partnering with Ouranos, CRIM, PCIC, CCCma, Natural Resources Canada, McGill and Concordia Universities.
Director of Platform Engineering (HPC and Cloud) - AstraZeneca, Mississauga ON CA
The Scientific Computing Platform (SCP) is a foundational capability for HPC and large-scale research computing solutions. Embedded within the Research D&A organisation it is central to analytics products focused on computational chemistry, imaging, multi-OMICs, structural biology, data science and AI, and is following a hybrid on-premises and public cloud strategy. The SCP team is accountable for the end-to-end delivery of high-performance analytics services, with an emphasis on augmenting the HPC experience. We combine modern HPC with a powerful DevOps stack and cloud-native technologies to power research and development at AstraZeneca.
The SCP Engineering Lead will both lead an agile engineering team and set the technical strategies to meet future scientific computing needs, where machine learning and AI are increasingly playing a major role in drug discovery at AstraZeneca.
Technical Advisor for High Performance Computing - National Research Council Canada, Mississauga ON CA
We are looking for a vibrant and dynamic technical advisor to join the Knowledge, Information, and Technology Services Branch (KITS) to support research centers with High Performance Computer (HPC) cluster maintenance and support. The advisor would be someone who shares our core values of impact, accountability, leadership, integrity and collaboration.
Senior Product Manager - High Performance Computing - 3M, Maplewood MN USA
3M is seeking a Senior Product Manager to join the Corporate Research Systems Lab (CRSL) to lead the creation and evolution of technology solutions that balance near-term business and technology needs and opportunities with long-term scalability, resilience, and low operating cost. The technical product manager establishes, owns, and evolves the unique function and business value of a complex product or set of products, relative to other technology products and services. A wide-ranging technical and domain expertise among business and research partners is needed to maximize the value of products and services through thoughtful data driven problem solving and functional influence among peer teams and organizations. As part of an agile team, you will enable applications in diverse markets including energy, manufacturing, personal safety, transportation, electronics, and consumer.
Team Leader Genome Research Infrastructure, German Human Genome-Phenome Archive (GHGA) - DKFZ, Heidelberg DE
The DKFZ is coordinating the German Human Genome-Phenome Archive (GHGA), a national data infrastructure dedicated to human omics data. GHGA is connected to major national initiatives, including genomDE, the German National Cohort (NAKO) and the National Center for Tumor Diseases (NCT). Internationally, GHGA serves as the German node both within the federated European Genome-Phenome Archive (EGA) and the European Genomics Data Infrastructure (GDI). The infrastructure will also play a key role in the national genomic initiative.
Statistical Sciences (CCS-6) Group Leader (R&D Manager 4) - Los Alamos Naional Laboratory, Los Alamos NM USA
The Statistical Sciences Group (CCS-6) has a long history of contributing to novel and high-impact projects in partnerships with world-leading scientists in a wide range of science, technology and engineering domains. CCS-6 performs research and development in the core foundational methodologies for accurate and defensible approaches to data analysis. CCS-6 develops methods for exploratory data analysis, design of experiments, and building models for prediction, classification, and uncertainty quantification. CCS-6 staff actively participate in interdisciplinary projects with application to nuclear deterrence, global security, and a variety of foundational science fields. The CCS-6 Group Leader is a member of the Computer, Computational, and Statistical Sciences Division leadership team and is responsible for technical leadership and management of the Statistical Sciences Group.
High Performance Computing Senior Manager, Institute of Computational Cosmology - Durham University, Durham UK
The Institute of Computational Cosmology (ICC) within the Department of Physics manages and operates one of the UK’s largest supercomputing facilities, COSMA, on behalf of DiRAC (www.dirac.ac.uk), a UK national facility. This role will be engaged with the continued operation of the COSMA system, including exciting new research and development projects, user support and operations. The applicant will liaise with DiRAC researchers in cosmology and other areas to understand their codes and develop expertise in the efficient running of these codes on supercomputing facilities. The applicant will have the opportunity to lead both software and hardware projects helping to shape the future of UK HPC. The applicant will also work with other support staff and other DiRAC technical support teams to deliver an effective service for users including routine and emergency maintenance activities. There will be opportunities for training to support the career development of the successful applicant, both within Durham and in collaboration with other DiRAC and/or industry partners.
Senior Product Manager, AI Foundations Research - CapitalOne, Various USA
As a Sr. Manager - Product on the AI Foundations Product team, you will be part of the team designing and delivering best in class ML and AI products for Capital One lines of business, data scientists, ML engineers and product teams. We’re passionate about experimenting and building products with next-generation technologies.
Cryo EM Facility Technical Director - University of Pennsylvania, Philadelphia PA USA
The successful candidate will provide management & supervision for the EMRL Core's various projects. Primary responsibilities include work in all phases of Electron Microscopy (EM), cryo Electron Microscopy (cryoEM), correlative cryogenic fluorescence light microscopy (cryoFLM) together with cryogenic electron tomography (cryoET) and cryo-FIB-SEM coupled with cryoET from sample preparation to expert operation & maintenance of the electron microscope.
Director, Digital Health & Informatics - Interior Health Authority, Various BC CA
The Director, Digital Health & Informatics is responsible for the nursing and allied health leadership that advances the utilization of technology enabled medical and clinical information management to improve quality outcomes for patients across clinical operations and team based care within clinical operations and the interdisciplinary health care teams across Interior Health, specifically providing nursing and allied staff leadership.
Engineering Lead – Scientific Computing Platform - AstraZeneca, Mississauga ON CA
The Scientific Computing Platform (SCP) is a foundational capability for HPC and large-scale research computing solutions. Embedded within the Research D&A organisation it is central to analytics products focused on computational chemistry, imaging, multi-OMICs, structural biology, data science and AI, and is following a hybrid on-premises and public cloud strategy. The SCP team is accountable for the end-to-end delivery of high-performance analytics services, with an emphasis on augmenting the HPC experience. We combine modern HPC with a powerful DevOps stack and cloud-native technologies to power research and development at AstraZeneca.
Associate Director, Research Software Engineering - Princeton University, Princeton NJ USA
The Research Software Engineering (RSE) Group, located institutionally in Princeton Research Computing but extending across campus, is hiring an Associate Director of Research Software Engineering. You will report to the Sr. Director of Research Software Engineering.In this position, you will build and lead a growing team of Research Software Engineers who provide dedicated expertise to researchers to create the most efficient, scalable, and sustainable research code possible to enable new scientific and scholarly advances. You will have the opportunity, and be encouraged, to bring new initiatives, technologies, and/or approaches to the RSE group and Princeton Research Software Community.
Head of Department of Informatics - Kings College London, London UK
Located in the heart of Central London on our Strand campus, the Department of Informatics within the Faculty of Natural, Mathematical & Engineering Sciences is a vibrant hub for interdisciplinary research and education of the highest quality. The department is in an exciting position, having grown by 30% over the last four years to 81 academic staff, with a student population of 1,500 undergraduates, 300 postgraduates and 150 PhD students. The successful candidate will have a substantial track record in research, teaching and professional achievement within an Informatics or Computer Science discipline, and evidence an aptitude and motivation towards inclusive leadership.
Program Manager, Digital Health and Discovery Platform - University Health Network, Various CA
The DHDP is a scalable, multi-use platform to digitally-enable national and international collaboration to advance next-generation precision medicine technologies. The DHDP is funded ($49M) from the Government of Canada’s Strategic Innovation Fund Stream 4 program. The DHDP will apply state-of-the-art data governance principles and technology to transform collaborative research and stimulate commercialization from home grown research discoveries.
We seek a highly motivated, collaborative, and knowledgeable professional who will be a key member of the team serving as the link between all operational areas of the organization. The Program Manager will contribute to the overall success of the DHDP through effective planning, execution and follow up of DHDPs organizational strategies, policies and administration.
Head, Data Science Training & Consultation - Stanford University, Stanford CA USA
This role is responsible for broad support for quantitative, computational, and algorithmic analysis of research data, including aspects of data management, analysis methods, workflow reproducibility, and ethical considerations. The Head of Data Science Training &Consultation continues a long tradition of university-wide service for data-driven research scholarship, while growing Research Data Services' capacity as a hub for digital and computational research support. In collaboration with other members of Stanford Libraries, the successful candidate will support this mission through campus-wide outreach efforts as well as the development of Library services that address the growing need for support of data science methods.
Director of Research Data Science - Stanford University, Stanford CA USA
This position is part of a new initiative incubated within Stanford Data Science, part of the Vice Provost for Research / Dean of Research. Stanford Data Science (SDS) is a dynamic and rapidly growing unit within the VP/Dean of Research. For more than four years, SDS has sought to advance data science and its application to all fields of study. Our community ranges broadly across all seven schools on campus, consisting of esteemed alumni, world class faculty, post-doctoral fellows and PhD students, dynamic staff and administrators. In realizing our mission, our staff are critical to supporting our organization’s goals and enabling Stanford faculty and students to accomplish their mission conducting cutting-edge research and innovation around how we learn from data, the tools we use, and the new methods needed to tackle the data-intensive future.
Centre Manager - Centre for Doctoral Training in Safe AI - University of York, York UK
We are seeking a pro-active, effective, and reliable individual to lead the operational delivery of the new UKRI UKRI AI Centre for Doctoral Training (CDT) in Lifelong Safety Assurance of AI-enabled Autonomous Systems (SAINTS). SAINTS is a prestigious new centre, funded to train 60 PhD students to become future leaders in Safe AI, whether in industry, academia, policy or regulation, and brings multiple disciplines together. Year 1 students will arrive in York in September 2024, and their application journey has already begun.
AI Institute Training Manager - University of Surrey, Surrey UK
An exciting opportunity for a Training Manager within the new Institute for People-Centred AI at the University of Surrey. We’re embarking on a new chapter for Artificial Intelligence, putting people at the heart of AI and augmenting human capabilities to deliver an inclusive and responsible force for good.
Director for Behavioral Health Research, AI Hub - New York University, New York NY USA
Reporting to the Executive Director of NYU McSilver, the Director for Behavioral Health Research will lend their subject matter expertise in behavioral health and supervise the Assistant Director of Research and AI Hub team, including graduate research assistants and consultants. The Director will also be responsible for study oversight, proposal development and implementation, manuscript preparation, coordination with the AI data analytics team, and strategic planning for the continued growth and expansion of research and evaluation projects. This role combines proficiency in behavioral health research to cultivate the creation of inventive solutions and visionary insights on youth behavioral health and public health challenges facing young people impacted by poverty. Additionally, the Director will spearhead initiatives to forecast youth suicide rates and behaviors by race, geography, income, and other demographic factors, with further groundbreaking public health research and interventions on the horizon. Additionally, the Director will work with staff and students on research projects and will be responsible for the pursuit, development, and expansion of research-oriented activities with strategic partners and funders.
Director of Illinois Computes - The National Center for Supercomputing Applications - University of Illinois, Urbana IL USA
NCSA is seeking a highly motivated leader to serve as the Director of Illinois Computes. The Director of Illinois Computes will provide coordination and leadership for the development and implementation of the Illinois Computes Program which includes existing long term investments from the University of Illinois System as well as the Urbana-Champaign campus. This will include integrating short-term tactical directions and long-range strategic directions of the program, establishing goals and timelines with program staff, and ensuring that staff and resources are directed efficiently toward achieving these goals within the specified timelines. In addition to directing the day-to-day operations of Illinois Computes, this position will provide oversight and management of assigned projects and/or programs operated by NCSA on behalf of the campus, national research communities, funding agencies, and other NCSA partner institutions
Research Engineering Manager - Thomson Reuters, London UK
As a Lead Research Engineer at Thomson Reuters Labs, you will be part of a global interdisciplinary team of experts. We hire engineers and specialists across a variety of AI research areas to drive the company’s digital transformation. The science and engineering of AI are rapidly evolving. We are looking for an adaptable learner who can think in code and likes to learn and develop new skills as they are needed; someone comfortable with jumping into new problem spaces; who enjoys directing and supporting the efforts of others.
High Performance Computing and Research Data Services Manager - Vassar College, Poughkeepsie NY USA
The High Performance Computing and Research Data Services Manager at Vassar College is a critical role in supporting research excellence and data management within the academic community. This position is responsible for managing and optimizing the high-performance computing (HPC) infrastructure and ensuring the efficient and secure management of research data and storage. The role involves collaborating with researchers, faculty, and students to enable cutting-edge projects while maintaining the integrity and accessibility of research data. The High Performance Computing and Research Data Services Manager reports to the Enterprise Systems team within Computing & Information Services and will work as a collaborative member of the Systems team to ensure operational continuity and support for HPC, Data Science, and other research-related computing and storage solutions. Computing & Information Services engages in cooperative efforts with the ACCAS Steering Committee, the Data Science & Society (DSS) Steering Committee, and various other academic departments to identify and implement priorities related to research computing.
Manager, High-Performance Computing - Columbia University, New York NY US
Reporting to the Sr. Director of Research Services; the Manager of High-Performance Computing (HPC) leads the HPC team and interacts extensively with leaders of other research support groups including the Office of the Executive VP for Research and other CUIT groups. Key initiatives to be accomplished by the incumbent include the advancement of a shared High-Performance Computing platform, the advancement of research computing in the cloud, assessment, and development of research tools for computational performance. Additionally, the Manager leads collaboration efforts with other central research support entities at Columbia University.
College of Engineering Research Computing Lead - University of Wisconsin at Madison, Madison WI USA
We need a a leader to help refine and grow our research computing services in the UW-Madison College of Engineering (CoE). This position will guide strategy to improve research computing efforts across the college, collaborating with technical colleagues, campus-level cyberinfrastructure efforts, and compliance professionals.
You will engage with researchers to understand their technology needs, translate that into service models, and drive implementation of those services. You will network with similar professionals across the UW-Madison campus to facilitate efficient use of computing services to advance research. In addition, you will work with wonderful colleagues to improve our readiness for audits around security, research integrity, and data protection.
Research Computing Support Lead (hybrid) - Northwestern University, Evanston, IL
As a Research Computing Support Lead, you will lead, develop, and supervise a team of staff and student computational consultants in delivering user support to researchers across Northwestern through tickets, consultations, workshops, and documentation. You will work with researchers and colleagues within and beyond Northwestern to understand emerging trends. You will oversee the development and implementation of necessary improvements and operational processes for support services to the meet growing and evolving research needs across the University.
Strong candidates for this role will have experience in supervising or coaching staff/students/interns, be experienced in leading or managing a service, and have a demonstrated record of building relationships.