About Me

I am a founding member at Ritual. I spend my time thinking about ML research and engineering problems for open-source machine learning models and their verifiability, privacy, and more.

Previously I was a founder of Socket where we used graph networks to associate crypto and social identity data. Before that, I spent time thinking about AI2050 and policy for robustness, interpretability, and data bottlenecks at Schmidt Futures.

I am interested in the robustness and applications of AI in the real world. I am in particular interested in human-AI, and AI-AI interactions over new domains and long horizons.

I worked on this topic in a few different ways.

1. policy at Schmidt Futures, EU AI policy (coauthor on Stanford HAI's input to the 2020 Whitepaper ), US national AI procurement policy (Federation of American Scientists expert profile).

2. research Stanford ML Group with Professor Andrew Ng, SAIL with Professor James Zou (interpretability, multi-agent competition)

3. engineering and research engineering. World Bank President's Office ( building models to deploy $2 billion USD to prevent famine casualties in Yemen and Somalia), Apple Sensing and Connectivity ML, Google Computer Vision, Infinitus (Healthcare AI), NASA Supercomputing (AI for solar flare)

4. lectured at Stanford on AI safety (designed and taught CS81si, AI interpretability and fairness). I am grateful to my faculty sponsors James Zou and Omer Reingold for this class.

Education. I obtained a Master's in Computer Science (AI track) and a Bachelor's in Mathematics at Stanford. I spent most of my time in college drinking coffee, going outside, and looking at my computer screen.

Things I like


I read and reread other people's blogs and longforms. I have written a few things in the past. Here's one that a few other people seemed to like (Top3 on HN). Below are some of my favourite technical and non-technical blogs.

Ben Kuhn's blog

Aaron's Swartz's old blogs

The Electric Typewriter, essay aggregator

Kevin Kelley's blogs

Math3Ma (Tai-Danae Bradley's blogs), Learning math intuitively.

Distill Pub for intuitive machine learning (now on hiatus)

Connected Papers


I am inspired by people working on hard problems. Starting a new project is one form of doing something hard. I enjoy chatting with people about new projects, particularly in AI or products that interact with the physical world. I am grateful to support founders in AI and beyond via investing.

Startup and venture communities have been a big part of my support system. Feel free to reach out if you want to chat about these things! The best way to reach me is by email.


If you create using a pseudonym and prefer interacting anonymously, please feel free to reach out with your pseudonym.

Research, Publications, Projects

Deepwind: Weakly supervised localization of wind turbines in satellite imagery
Sharon Zhou*, Jeremy Irvin*, Zhecheng Wang, Eva Zhang, Jabs Aljubran, Will Deadrick, Ram Rajagopal, Andrew Ng
NeurIPS'19 | 33rd Conference on Neural Information Processing Systems
PDF 2019

Competing AI: How does competition feedback affect machine learning?
Tony Ginart, Eva Zhang, Yongchan Kwon, James Zou
AISTATS'21 | International Conference on Artificial Intelligence and Statistics
PDF 2021

Latent Actor-Critic with Intrinsic Motivation and Skill Hierarchy
Ademi Adeniji*, Eva Zhang*
PDF 2020

Hypothesis Formation Using Laterality-Tagged Survival Data Analysis
Eva Zhang, Sarah Poole, Nigam Shah
NIPS Workshop
poster 2018

A National Framework For AI Procurement
Eva Zhang, Grant Gordon, Katie Jonsson
Federation of American Scientists
PDF 2021

Technology Modernization Fund: Learning from Tech Investment Funds
Tech Talent Project
PDF| Note: Not a coauthor (acknowledgement only). Linking because it's a cool report and you should check it out. 2021

Input on the European Commission White Paper “On Artificial Intelligence – A European approach to excellence and trust
Stanford HAI
Stanford HAI · Jun 15, 2020
PDF 2020