Category 1 -- Highest Priority
Teams at the intersection of AI and financial markets, specifically focused on using ML/AI to identify investment opportunities. Most directly aligned with the AI Warren Buffett mission.
Bay Area focus: 3 companies in the main section. 7 non-Bay Area companies archived below for reference.
Numerai Bay Area
Key People
Richard Craib
Product: Crowdsourced AI Hedge Fund
- Provides encrypted, obfuscated financial data to thousands of data scientists globally -- nobody knows what stocks they are predicting
- Data scientists build ML models on encrypted data and submit predictions weekly
- Best models are staked with NMR (Numeraire cryptocurrency) -- skin in the game
- Models that perform well earn NMR; poor models lose their stake
- Numerai ensembles the best models into a meta-model for actual trading -- "wisdom of the ML crowd"
- Uses homomorphic-style encryption so nobody knows what stocks they are predicting, preventing insider trading and information leakage
- Thousands of active participants globally contributing models
Compensation (ML Engineer)
- Small team (~30), estimated $200K-$350K total comp for ML Engineers
- NMR token upside as additional compensation component
- Startup-level equity + crypto token economics
Why It Matters
Crowdsourced "wisdom of the ML crowd" -- thousands of models combined into a meta-model. Innovative approach to incentive alignment via crypto staking. Direct exposure to ML for finance at scale. The encryption approach prevents overfitting to known securities, which is a key insight for building robust financial ML models. Small team means enormous impact per person.
Path to Entry
- Participate in the tournament first to demonstrate skill and interest
- ML Engineer, Research Engineer roles
- Heavy focus on ML depth and model evaluation under distribution shift
- Experience building scalable ML pipelines from Meta is directly relevant
- Small team means direct access to Richard Craib and core problems
Tracking
Key People
Michael Kharitonov
Jon McAuliffe
Product: ML-First Systematic Trading
- Unlike traditional quant funds that use hand-crafted signals, Voleon trains deep learning models directly on financial data to generate trading signals
- ML models process tick-level market data, financial statements, news/text data, and alternative data
- Models output portfolio weights that are directly executed -- no human traders in the loop
- Fully automated end-to-end pipeline from data ingestion to trade execution
- ML is not a tool augmenting human traders -- it IS the investment strategy
- One of the most pure-play ML investing firms in existence
Compensation (ML Engineer / Researcher)
- $300K-$600K+ total comp for ML Engineers/Researchers
- Quant fund compensation with performance-based bonus
- PhD preferred for research roles
Why It Matters
One of the most pure-play ML investing firms. ML is not a tool augmenting human traders -- it IS the investment strategy. Berkeley-based with deep academic pedigree from UC Berkeley CS and statistics. $8B+ AUM proves the ML-first approach works at scale. For someone building an AI Warren Buffett, this is where you learn how to make ML directly generate investment decisions end-to-end.
Path to Entry
- Strong ML/statistics background required
- PhD preferred for research roles
- Quantitative Research and ML Engineering roles
- Meta ML experience is directly relevant -- scaling, distributed training, model optimization
- Berkeley location is ideal for Bay Area
Tracking
Fintool Bay Area
Key People
Founding Team
Product: AI Financial Research Assistant
- Uses LLMs fine-tuned on financial documents -- SEC filings, earnings transcripts, 10-Ks, 10-Qs
- Answers complex financial questions with precise citations (page/line from source documents)
- Example: "What was Apple's revenue growth in Q3 vs Q2?" returns precise answer with page/line citations from the 10-Q
- Uses RAG (Retrieval Augmented Generation) over a proprietary database of financial filings
- Focuses on accuracy and hallucination reduction -- critical for financial use cases where wrong numbers can be costly
- Built specifically for the financial domain where precision matters more than creativity
Compensation (ML / Full-Stack Engineer)
- Startup comp, estimated $200K-$350K total
- Early-stage equity with significant upside potential
- YC-backed with strong fundraising trajectory
Why It Matters
Directly applying LLMs to financial analysis -- exactly the type of tool an AI Warren Buffett would need. The focus on precise, cited answers from financial documents is the foundation layer for any AI investment system. YC-backed, SF-based, small team = high impact per person. Solving hallucination reduction for financial use cases is one of the hardest and most important problems in applied LLMs.
Path to Entry
- ML Engineer, Full-Stack Engineer roles
- Small team means generalist skills valued alongside ML depth
- RAG, LLM fine-tuning, and information retrieval experience highly relevant
- Meta ML experience provides strong signal for a small startup
- SF location aligns perfectly
Tracking
Archived -- Not Bay Area (7 companies)
▶Reflexivity (Toggle AI) New York, London, Tokyo
Why Archived
HQ in New York with offices in London and Tokyo. No Bay Area presence. Founded by Jan Szilagyi (former $3B+ hedge fund manager). Backed by Stanley Druckenmiller. Integration with Interactive Brokers for distribution at scale. Named after Soros's reflexivity concept -- market participants' biases influence fundamentals which influence biases, creating feedback loops. Strong concept but requires East Coast relocation.
Two Sigma New York, NY
Why Archived
HQ in New York City. Technology-driven systematic investing using ML and distributed computing. Founded by David Siegel (PhD MIT CS) and John Overdeck (IMO silver medalist). Massive alternative data infrastructure -- ingests satellite imagery, social media, financial filings, web scraping, IoT data. Engineering culture closer to Google/Meta than Wall Street. Gold standard for AI-driven investing but requires NYC relocation.
Bridgewater Associates Westport, CT
Why Archived
HQ in Westport, Connecticut. Founded by Ray Dalio. AIA Labs initiative is building "Artificial Investment Associate" to codify investor thinking into AI systems -- most directly aligned effort with AI Warren Buffett concept. Radical transparency culture. Dalio's Principles framework is essentially "training data" for an AI investor. Incredible mission alignment but requires Connecticut relocation.
AlphaSense New York, NY
Why Archived
HQ in New York City. AI-powered market intelligence platform using NLP to search across financial documents. Smart Synonyms technology understands financial jargon and context. Sentiment analysis detects changes in management tone. Founded by Jack Kokko and Raj Neervannan. 2,000+ enterprise customers including major banks and hedge funds. Strong product but NYC-based.
Kensho (S&P Global) Cambridge, MA / New York
Why Archived
HQ in Cambridge, MA with NYC office. Acquired by S&P Global for $550M. Builds NLP tools for financial data: Scribe (speech-to-text for earnings calls), NERD (financial entity recognition), Link (entity linking across data sources), Extract (automated data extraction from filings). Founded by Daniel Nadler (Harvard PhD). Access to world's richest financial dataset via S&P. East Coast only.
Danelfin Barcelona, Spain
Why Archived
HQ in Barcelona, Spain. AI Score (1-10) rating for stocks based on technical, fundamental, and sentiment analysis. Small team (1-10 employees). Key differentiator is explainability -- shows which indicators drive the score with Technical/Fundamental/Sentiment sub-scores. Interesting product but requires international relocation.
Kavout Seattle, WA
Why Archived
HQ in Seattle, WA. AI investment analytics platform with proprietary "Kai Score" (1-9) stock rating analyzing 200+ factors across fundamental, technical, and sentiment dimensions. Small team, serves both retail and institutional clients. Directly building an AI stock picker. Seattle location means not Bay Area.