To announce that you are nominating yourself please respond to this post by November 4th. We encourage all nominees to briefly summarize their experiences with Stan and their goals for the Stan Governing Body to assist the electorate in informing their votes. Feel free to add links to any content you think is relevant.
If you know someone who you think should run for the SGB then we encourage you to contact that person and suggest that they nominate themselves.
I am nominating myself for the Stan Governing Body. I have a strong interest in Stan development and advocating its use, particularly for pharmacometric applications. My qualifications and priorities for the SGB are summarized below along with a brief bio.
Sincerely,
Bill Gillespie
William R (Bill) Gillespie, Ph.D.
VP Strategic Modeling & Simulation
Metrum Research Group (https://metrumrg.com)
107 Duryer Court
Cary, NC 27511 billg@metrumrg.com
Qualifications & experience relevant to SGB functions
Stan user since 2014
Internationally recognized in the pharmacometrics community as a leading expert, advocate and educator in the use of Bayesian pharmacometric data analysis.
Presented several workshops on Bayesian pharmacometric data analysis using Stan
Getting Started with Bayesian PKPD Modeling Using Stan: Practical use of Stan and R for PKPD applications
ACoP6, Arlington, VA, 8 October 2015
PAGE, Lisbon, 10 June 2016
ACoP7, Bellevue, WA, 23 October 2016
PAGE, Stockholm, 14 June 2019
Advanced Use of Stan, rstan and Torsten for Pharmacometric Applications
PAGE, Budapets, 10 June 2017
ACoP8, Fort Lauderdale, FL, 15 October 2017
PAGE, Montreux, 29 May 2018
Bayesian analysis of categorical, count and time-to-event data with Stan, rstanarm and Torsten
ACoP9, San Diego, 7 October 2018
Provide online training resources in the use of Stan for pharmacometric applications
A Brief Introduction to Bayesian Modeling Using Stan
In collaboration with Stan developers at Columbia University successfully obtained an STTR grant from ONR in 2016 entitled “Fast and Flexible Differential Equation Model Fitting with Application to Pharmacometrics” (https://www.sbir.gov/sbirsearch/detail/1587953). I am a co-PI with Andrew Gelman. This supports the development of Torsten as well as new components of Stan, e.g., the algebraic solver.
My priorities as a SGB candidate
I am committed to the use, development and promotion of high quality open source, publicly licensed software for scientific applications. Consistent with that, I want to do my part to support the Stan development effort and participate in efforts to assure it’s continuation as the pre-eminent platform for Bayesian data analysis.
Continue and expand the active Stan development effort.
Promote effective use of Stan via meetings, e.g., StanCon, training and publications.
Obtain the resources needed to accomplish those efforts.
Brief Biography
Education
1987 Ph.D. Pharmacy, The University of Iowa, Iowa City, Iowa.
1980 M.S., Pharmacy, The University of Michigan, Ann Arbor, Michigan.
1976 B.S., Pharmacy, Wayne State University, Detroit, Michigan.
For more than 30 years I have been involved in the development and application of pharmacometric methodology for enhancing drug treatment, development and regulation. From 1980 to 1984, I was a Medical Research Associate with the Biopharmaceutics Unit of The Upjohn Company. From 1987 to 1993, I was Assistant Professor of Pharmaceutics at The University of Texas at Austin, College of Pharmacy. In 1993 I joined the Center for Drug Evaluation and Research, U.S. Food and Drug Administration where he served as the Associate Director for Scientific Affairs, Office of Clinical Pharmacology and Biopharmaceutics. From 1997 to 1999 I was the Vice President for Pharmacokinetic Research and Development at GloboMax LLC. I then joined the Pharsight Corporation as a Senior Scientific Consultant from 1999 to 2007. I joined Metrum Research Group in 2007 where I am currently employed as VP of strategic Modeling & Simulation. I have published over 60 manuscripts and was the 2014 recipient of the International Society of Pharmacometrics Innovation Award.
My research interests include theoretical and computer analysis of pharmacokinetic and pharmacodynamic systems. My recent efforts have concentrated on the use of Bayesian modeling and simulation to optimize decision-making in clinical drug development and treatment. I am seeking to make Bayesian modeling and simulation methods more accessible via training programs and development of software tools, e.g., Torsten (https://github.com/metrumresearchgroup/Torsten) and BUGSModelLibrary (https://bitbucket.org/metrumrg/bugsmodellibrary/wiki/Home).
I am Leah Comment, and I would like to nominate myself for the SGB.
My background:
I graduated in May 2019 with a PhD in Biostatistics from Harvard. You may have seen me presenting at StanCon Asilomar (2018) or StanCon Cambridge (2019); I also helped to teach a Stan short course at ODSC East 2018. I have been a Stan convert since attending my very first StanCon back in 2017, and two of my three dissertation papers involved writing Stan models. I am now Associate Director of Decision Sciences at Foundation Medicine, a molecular insights company focused on cancer.
In terms of my background and current affiliations, I represent the intersection of a few Stan constituencies:
I am a statistician actively working to develop new Bayesian methods (topic areas: causal inference for personalized health care and innovative clinical trial design).
I work in the biotech industry. Some of my research involves products and therapies overseen by regulatory authorities, while others do not.
Through my experience teaching and consulting for public health students outside the biostatistics department, I understand the need for basic, user-friendly Bayesian analogs to frequentist approaches. A Stan language that is too focused on the cutting edge will miss the opportunity to improve today’s scientific research with Bayesian methods that already exist.
As I build out a new function at a company that has not previously used Stan, I care about how to identify good Stan programmers and how to get non-users up to speed efficiently.
I am personally and professionally interested in how to communicate uncertainty around high-stakes decision making.
With that said, my goal is not to represent only these groups. I also want to give back to the wonderful Stan community which has been so generous with its time and expertise.
My priorities for the Stan Governing Body are as follows:
(1) Build a scalable Stan training program
Scope and encourage development of a Stan multi massive open online course (MOOC).
Gauge industry interest in supporting open training initiatives (see (2) below).
(2) Identify industry players, and survey them to assess current use and future needs
Are there areas of overlapping need where companies can jointly fund a solution to be added to the Stan code base?
Are there best practices that can be shared more broadly without jeopardizing intellectual property?
What are the common needs surrounding (a) user training and (b) deployment of Stan models on “industry-scale” data?
(3) Foster an inclusive, vibrant Stan community
Support student and postdoc access to StanCon through travel awards.
Identify and encourage connections for common interest groups (e.g., professors teaching Bayesian methods with Stan, users of rstan Docker containers in AWS Batch).
Connect local Bayesian meetups to qualified, engaging teachers of the Stan language.
I am nominating myself for the next SGB because I believe that I can be an effective communicator and facilitator between various important groups within the Stan community:
I have been a Stan developer since 2014 and have good relationships with developers in all areas of Stan development. At the same time I also spend a lot of my time interacting with Stan users on the forums and when traveling to teach and promote Stan. From these experiences I am in touch people using Stan for so many different purposes in so many different places, and I have learned a lot about the challenges that face potential new users when they first encounter Stan.
I work in academia (focusing on applied Bayesian statistics), but I also have a lot of experience helping companies use Stan in various industries and helping other organizations (e.g., non-profits, governments) use Stan for many different important applications.
As a member of the current provisional SGB, I can help smooth the transition to the next group so the new SGB can get off to a fast start. I have also learned a lot from what went well and what didn’t go well for the provisional SGB and am very motivated to use that experience to help the next SGB succeed.
Overall, I think the role of the SGB should be to provide transparent oversight for the project and to empower people who care about Stan to help the project move forward. There are a lot of important challenges facing the next SGB: project governance issues (e.g. defining a new electorate and developing a code of conduct), community building, and maintaining and improving our open source software, among others. The SGB alone can’t solve these issues. My priority will be to make sure that the SGB works well together and in concert with other people who have the experience, skills, and energy we will need to rely on in order for Stan to flourish.
I’m excited to see the nominations we already have and I would be honored to work with this group of people to help guide Stan forward.
My name is Susana Marquez and I am nominating myself for the Stan Governing Body.
My background:
I currently live in New York and work as a Manager and Data Scientist at The Rockefeller Foundation. Prior to my current role, I was a data scientist at BuzzFeed, and a statistical geneticist at The Feinstein Institute for Medical Research in New York, where I used Stan to estimate the effects of different treatments on proteins to develop a test for the early detection of endometriosis. Like many of the other candidates, I’ve attended previous Stan conferences, including the first in NYC (2017) and StanCon Asilomar (2018). Additionally. I’ve been actively involved with the community that gathers in NYC for the meet-up events. I hold a Graduate Diploma in Statistics and Stochastic Processes from the University of Melbourne (Australia), and a BA (H1) from the University of Guadalajara (Mexico).
I work in philanthropy, where my job consists of assessing the quality of technical pitches. I am familiar with many legal, operational and communications challenges that nonprofit organizations face, and I think I can bring some of these skills and advice as an SGB candidate.
My interest and priorities as an SGB candidate:
a) Identify potential funding partnerships
Identify funding mechanisms that can become sustainable in the medium-term.
b) Promote the use of Stan (and open-source software more generally)
Often, companies worry about future maintenance of Stan code if their staff leave the company. Identifying and learning from cases where non-users were trained to be able to give continuity to projects can help develop a strategy to promote use in industry.
c) Advance diversity in the Stan community
There are several minority students and professionals who routinely use Stan and often travel to non-English/non-Western countries to provide training and promote the use of Bayesian techniques. I believe these folks should be more actively recognized and supported.
I am a member of R-Ladies Global and I believe this organization has been very successful in creating a welcoming environment in the R community. I think collaborating with similar groups can help foster a more inclusive community.
Disclaimer: My husband, James Savage, recently made a grant to Stan as he is also interested in seeing this project grow and he has been an active member since the early days of Stan. Before my nomination, I consulted with several members of the community and some current SGB members to make sure there are no conflicts of interest.
Thank you,
Susana Marquez
Manager and Data Scientist
The Rockefeller Foundation
suemarquez at gmail dot com
I would like to nominate myself for the upcoming Stan Governing Body (SGB). From my experience I find Stan to be an extremely useful statistical model building tool both in industry and academia. I would love to help facilitate the growth of the project and foster a wider adoption of Stan in industry. Below I have outlined my goals for the SGB as well as my background.
Goals for the SGB
My goals for the SGB include improving the software development process and dealing with what I perceive to be a PR problem with Stan (in some industries).
Software Development
Provide more transparency and coordination on the software development process, possibly through the introduction of a Technical Working Group. Working with engineers in industry I’ve seen projects struggle due to
lack of coordination and scope creep, and
neglecting the importance of prioritizing and resolving user issues/requests.
Encourage (and make it easier for) new developers to join and actively contribute to Stan.
Resolve Stan’s PR Problem
Expose Stan to a more diverse range of industries where the modeling language would be useful. I’ve struggled to do this in the industries that I’ve worked in but have learned from the issues I’ve faced. (The goal would be to do this while maintaining Stan’s presence in existing industries and in academia.)
Improve our web-based informational resources (i.e. website and documentation).
Make it more accommodating to new users (particularly those in industry).
Improve the discoverability of user and developer resources.
Background and Familiarity with Stan
I am currently working in industry as a Senior Data Scientist at PlayStation in San Mateo, California (previously at the National Basketball Association).
I have been working on/with Stan since 2016.
Contributed modeling functions and developer onboarding documentation to the rstanarm R package.
Worked on one of the MRP election papers with Andrew Gelman.
Was the Teaching Assistant for Ben Goodrich’s Bayesian Statistics for the Social Sciences class at Columbia University.
Have contributed case studies to the Stan website.
Have attended the first two StanCons and semi-regularly attended the Stan Meetups in NYC.
I am nominating myself for the next SGB. I believe that I can be a effective bridge between cognitive scientists, psycholinguists, psychologists, and other scientists with the Stan project:
My background
I’m a cognitive scientist/psycholinguist about to start an assistant professorship in department of Cognitive Science and AI at the University of Tilburg. I have been a Stan user since 2013, and I have been a “heavy” user of Stan in my own research: I’ve used it to analyze experimental work and for developing cognitive models, and I’ve presented some of my work in two Stancons (Stancon2017 and Stacon2018 Helsinki). My complete CV is in my website.
I have spent considerable amount of time introducing Bayesian statistics and Stan to non-statisticians:
I authored and co-authored several tutorials to introduce rstan, rstanarm and brms.
I have taught Bayesian statistics using (r)Stan and brms in Master courses, workshops and summer schools.
I’m currently working (together with Shravan Vasishth and Daniel Schad) on a book that introduces Bayesian statistics to cognitive scientists using Stan and brms.
Limitations of Stan (as I see them) and my priorities as a SGB candidate
Even though Stan has a lot of documentation, the vastness of might be actually preventing many scientists to start using Stan. In addition, because Stan is growing so fast, many (cognitive) scientists feel that best practices are a moving target (e.g., non-centered parametrization, marginalizing discrete parameters, recommend priors), and feel frustrated. I think that given how useful Stan is to develop models is also quite surprising the relative absence of, for example, cognitive modeling and neuroscience in the case studies and in the Stancon presentations.
My priority as a SGB member would be the following:
making the use of Stan more accessible by coordinating the creation of better documentation, website, (installation) guides, and case studies that grow along with the incredible work in development,
broadening the community by making Stancon and case studies more friendly to absent disciplines, and
incentivizing and facilitating the contribution to Stan and its ecosystem from the community
Thank you to everyone who nominated themselves! Everyone in the current electorate should now have received an email with a link to the online ballot (via https://heliosvoting.org/). Voting ends on Nov 11th.