Ethan Fast

I work across startups, machine learning, biology, programming tools, and public writing. My most recent work is on Vcreate, a company I co-founded in 2020 that builds high-throughput platforms for T-cell receptor discovery; our work has led to multiple SBIR awards, patent filings, and machine learning systems for immunology.

Before that I co-founded Nash, where we use state channels and MPC to make self-custody feel closer to centralized finance UX. The platform now lets users in Europe manage traditional bank accounts and self-custodied crypto assets in one place. We also issued the first regulated European security token.

I have a PhD in Computer Science from Stanford, where I worked on human-computer interaction, language, and machine learning, including Best Paper award-winning systems and papers at CHI, EMNLP, AAAI, UIST, and Nature Biotechnology.

Public Work

  1. writing

    Distilling a Tiny Model for Fast Interpretability, Substack.

    I trained a ~40M-parameter student model to approximate Gemma 4B representations, and found that it preserved enough structure to support SAE-based analysis.

    Code: fast-ai-detector, GitHub.

  2. research

    Rapid design and validation of highly specific binders against MHC-peptide targets, Journal for ImmunoTherapy of Cancer. Ethan Fast, Manjima Dhar, Millicent Ku, Binbin Chen; Vcreate.

  3. grant

    T-cell receptors for treating pediatric acute myeloid leukemia, NIH/HHS SBIR Phase I. Vcreate. $355,000.

  4. patent

    Methods for finding novel T-cell receptors, PCT patent filing. Vcreate; Binbin Chen, Ethan Fast, Manjima Dhar.

    Published Feb 2025.

  5. grant

    High-throughput and antigen-specific TCR discovery for solid tumor treatments, NSF SBIR Phase II. Vcreate. $1,000,000.

  6. research

    TAPIR: a T-cell receptor language model for predicting rare and novel targets, bioRxiv. Manjima Dhar, Ethan Fast, Binbin Chen; Vcreate.

    Platform: TAPIR web service.

    Press: TCRs and AI: the future is now, Nature Reviews Immunology.

    Usage: 2M+ TCRs processed, with users from hundreds of institutions.

  7. grant

    High-throughput antigen-specific TCR discovery through engineered macrophages, NSF SBIR Phase I. Vcreate. $256,000.

  8. patent

    Compositions and Methods for Making Novel T-Cell Receptors, U.S. patent filing. Vcreate; Binbin Chen, Ethan Fast.

    Published Mar 2023; issued Apr 2026.

  9. grant

    Computational algorithm to predict interacting MHC alleles from TCR sequences, NIH/HHS SBIR Phase I. Vcreate. $256,581.

  10. research

    Improving Security for Users of Decentralized Exchanges Through Multiparty Computation, IEEE International Conference on Blockchain 2021. Robert Annessi, Ethan Fast; Nash.

  11. company

    Co-founded Vcreate, focused on high-throughput T-cell receptor discovery using novel assays and machine learning.

  12. open source

    OpenLimits, a Rust high performance cryptocurrency trading API with support for multiple exchanges; Nash.

  13. writing

    Can artificial intelligence help us design vaccines?, Brookings. Ethan Fast, Binbin Chen.

  14. research

    How to train your learners: Reinforcement learning for the scheduling of online learning activities, CHI 2020. Jonathan Bassen, Bharathan Balaji, Michael Schaarschmidt, Candace Thille, Jay Painter, Dawn Zimmaro, Alex Games, Ethan Fast, John C. Mitchell.

  15. research

    Potential T-cell and B-cell Epitopes of 2019-nCoV, bioRxiv. Ethan Fast, Russ B. Altman, Binbin Chen.

    Patent: cited in later SARS-CoV-2 vaccine patent filings.

  16. writing

    The Secret for the Best UX? Controlling Your Own Data, Observer.

  17. research

    Predicting HLA class II antigen presentation through integrated deep learning, Nature Biotechnology. Binbin Chen, Zining Khodadoust, ..., Ethan Fast, Russ B. Altman.

  18. degree

    Completed PhD at Stanford: Capturing Human Behavior and Language for Interactive Systems.

  19. research

    Iris: A Conversational Agent for Complex Tasks, CHI 2018. Ethan Fast, Binbin Chen, Julia Mendelsohn, Jonathan Bassen, Michael S. Bernstein.

    Code: iris-agent, GitHub.

    Essay: A Conversational Agent for Data Science, HackerNoon.

  20. talk

    Faculty/job talks on capturing human behavior and language at UC Berkeley, MIT, Harvard, and others.

  21. company

    Co-founded Nash, a non-custodial exchange and crypto asset management platform.

    State channels and MPC systems to provide good trading UX without centralized custody.

    Essay: Security tokens: How Nash views regulation in the new digital economy, Medium / Nash.

    Press: CryptoSlate interview on self-custody UX; CNBC Advancements feature.

  22. research

    Lexicons on Demand: Neural Word Embeddings for Large-Scale Text Analysis, IJCAI 2017. Ethan Fast, Binbin Chen, Michael S. Bernstein.

  23. open source

    Neon Wallet, light wallet for managing digital assets on the NEO blockchain.

    Created the original wallet and passed control to new maintainers in 2018; current project site.

    Most popular wallet on the NEO blockchain for several years, 100k+ users.

  24. grant

    Measuring Public Perception of Stories in the News, Brown Institute Magic Grant. Ethan Fast, Binbin Chen.

  25. research

    Long-term Trends in the Public Perception of Artificial Intelligence, AAAI 2017. Ethan Fast, Eric Horvitz.

    Essay: Long-term Trends in the Public Perception of Artificial Intelligence, HackerNoon.

  26. writing

    What’s in an Inauguration Speech? A Lexical Analysis via Empath, HackerNoon.

  27. writing

    The Best Fiction I Read in 2016, Medium.

  28. research

    Identifying Dogmatism in Social Media: Signals and Models, EMNLP 2016. Ethan Fast, Eric Horvitz.

  29. research

    Meta: Enabling Programming Languages to Learn from the Crowd, UIST 2016. Ethan Fast, Michael S. Bernstein.

  30. writing

    On Introductions and Research Papers, Medium.

  31. writing

    A Student’s Guide to PC Meetings, Medium.

  32. award

    Best Paper, CHI 2016. For Empath: Understanding Topic Signals in Large-Scale Text.

  33. award

    Best Paper Honorable Mention, CHI 2016. For Augur: Mining Human Behaviors from Fiction to Power Interactive Systems.

    Press: Globe and Mail, Engadget, CBC Spark, Singularity Hub.

  34. research

    Empath: Understanding Topic Signals in Large-Scale Text, CHI 2016. Ethan Fast, Binbin Chen, Michael S. Bernstein.

    Code: empath-client, GitHub.

    A tool for analyzing text across lexical categories, widely used in computational social science work.

  35. research

    Augur: Mining Human Behaviors from Fiction to Power Interactive Systems, CHI 2016. Ethan Fast, William McGrath, Pranav Rajpurkar, Michael S. Bernstein.

  36. research

    Shirtless and Dangerous: Quantifying Linguistic Signals of Gender Bias in an Online Fiction Writing Community, ICWSM 2016. Ethan Fast, Tina Vachovsky, Michael S. Bernstein.

  37. research

    Text Mining Emergent Human Behaviors for Interactive Systems, CHI Extended Abstracts 2015. Ethan Fast, Pranav Rajpurkar, Michael S. Bernstein.

  38. research

    Software mutational robustness, Genetic Programming and Evolvable Machines. Brian J. H. H. Langdon, Mark Harman, Yue Jia, Ethan Fast, Westley Weimer, Stephanie Forrest.

  39. research

    Emergent, Crowd-scale Programming Practice in the IDE, CHI 2014. Ethan Fast, Daniel Steffee, Lucy Wang, Joel R. Brandt, Michael S. Bernstein.

  40. award

    NSF Graduate Fellowship.

  41. research

    Crowd-scale Interactive Formal Reasoning and Analytics, UIST 2013. Ethan Fast, Colleen Lee, Alex Aiken, Michael S. Bernstein, Daphne Koller, Eric Smith.

  42. research

    Software Mutational Robustness: Bridging The Gap Between Mutation Testing and Evolutionary Biology. Mark Harman, Yue Jia, William B. Langdon, Ethan Fast, Westley Weimer, Stephanie Forrest.

  43. open source

    TypedJS, experimental JavaScript fuzz testing with type annotations.

  44. company

    Co-founded Proxino, a Y Combinator-backed JavaScript analytics and error reporting company.

    Press: TechCrunch covered Proxino.

  45. research

    Designing better fitness functions for automated program repair, GECCO 2010. Ethan Fast, Claire Le Goues, Stephanie Forrest, Westley Weimer.

  46. open source

    Gajure, a Clojure framework for genetic algorithms.

  47. writing

    Early essay archive: Go Meta, The Signal is not the Goal, Old Palo Alto, Reciprocity and the Status Quo, Ask the Question, See Things as They Are, Code that Writes Code, More on Coderacy and Mental Models, The Aspirational Polymath, Programmers Can Be Artists, Don't Jump off the Cliff with a Parachute, You're Afraid to Be Anything, The Rise of Coderacy, Website A/B Testing and Unknown, Unknowns, Programming, Education, and Rationality, Deconstructing a Successful YC Application, Minimalist Running, The Artful Mastering of Goals, Advice for Undergrads, State Your Purpose, Metaprogramming Ruby, The Allure of the Asymmetrical, On Parallelism in Gajure, A Genetic Algorithm Framework in Clojure.