Personal

I live in New York City.

• Alberta, Canada is home (and where I grew up), but I was born in India
• I am interested in quantitative finance, deep learning, modelling biological systems, and supercomputers
• Philosophers I admire include: Francis Bacon, Thomas Aquinas, Immanuel Kant, and René Descartes
• In another life, I would have pursued a career in storytelling — and love literature that explores worldbuilding, games within games, and sci-fi (e.g. Tolkien, Cixin Liu, R.R. Martin, Herbert)
• I aspire to explore every country on Earth, and am grateful to be 21% of the way there

Background & Work

I am currently serving a 15 month non-compete with Citadel Securities, the quantitative trading firm.

• During this time, I am a senior advisor on AI to a $1B+ AUM venture capital firm based in NYC
• My priorities during this sabbatical are to catch up on readings at the deep learning frontier — in particular, the latest results in sequence modelling, multi-agent systems, and DNA foundation models
• Please do not reach out with employment opportunities — I will have to politely decline
• Previously, I worked at Citadel Securities, in the Special Projects group based out of NYC. Much of what I worked on remains highly confidential and proprietary to the Firm
• Before that, I worked at Microsoft Research in Seattle
• I studied Computer Science and Economics at Harvard University
• At Harvard, much of my time was spent working at the Broad Institute of Harvard/MIT, which culminated in a senior thesis supervised by Professor Finale Doshi-Velez on modelling healthcare information
• When I was younger, I was very involved in the national science fair, olympiad, and debate circuits across Canada

Interests & Projects

On the highest level, I am interested in building prediction machines — systems that learn the underlying distribution of information in a given setting through data, and correctly sample from this distribution at inference time.

Quantitative Biology:

• A grand challenge I am personally interested in is modelling a virtual cell
• I am inspired by the AlphaFold results, which uncovered the underlying manifold representing the distribution of proteins, and believe similar manifolds exist throughout the cell (DNA, lipids, etc) due to evolutionary pressure from nature

Finance & Capital Markets:

• I am highly interested in the quantitative trading industry, and how markets have evolved concurrently with progress in computing. I am interested in new frontiers for market structure, exchange matching algorithms for improved price discovery, and auction design
• I've enjoyed literature on the history of commodity markets — starting with the Dojima rice exchange in Japan (the first modern futures market) to the development of lit oil markets — and how centralizing trade among these markets has accelerated global development
• I believe prices — and improvements in technology that lower the price or increasive the produtivity per unit of a good— tell the story of economic growth (and often social/political stability of nations) better than any other variable in the long-run
• In finance outside of trading and markets, I also have an interest in the deal-making business — specifically private equity and venture investing

Computer Science:

• Every few years, I try to re-read the classics of computer science, and have built a sheet of the key papers that defined the field
• Recently, my interests have been in deep learning, specifically in new architectures for data-efficient learning, long-context sequence modelling, and multi-agent systems
• The fact that computers can now build an increasingly accurate representation of information in the universe — through learning the underlying distribution of training data — is in and of itself remarkable
• The fact that we can sample from this distribution — both to understand reality and potentially produce net-new knowledge — makes these systems highly desirable
• A promising research direction I am tracking is teaching models how to accurately action tools: both in the digital world (software, operating systems, etc) and by extension the physical world (microscopes, vehicles, etc), with applications in AI for automating scientific research
• I believe the demand for compute follows Jevon's paradox — and that recent results with search and planning of AI systems illustrates the demand for supercomputing will stand the test of time
• On a systems level, I am interested in new orchestration engines for managing compute containers at scale, programming languages centered around tensor artifacts with the performance latency of C++, and the relationship between datacenters and power markets

Business & Investing:

• In a limited capacity, I will cut small checks into startups going after frontiers I am passionate about, especially when I have a pre-existing relationship with the founder
• A new venture I am interested in is the creation of a new market structure for global compute, especially GPU compute time, to treat it as a true commodity with price discovery. I believe this will help accelerate progress in AI significantly

Miscellaneous:

• At Harvard, I served as the President of HPAIR, and organized a summit with Global CEOs, Heads of State, and World Leaders that Business Insider called the "Davos of Harvard"
• I was once roped into a side-project to detect exoplanets from astronomy data using neural networks for a summer

Contact

Please email me directly if you would like to get in touch. I do not use most social media platforms, so this homepage serves as my voice on the Internet.

• zeel@zeelmpatel.com via email
• @zeelmpatel on Twitter