I asked people: “which data companies should an ambitious future founder join?”
Human expertise for frontier AI, from model training and evaluation to useful work.
Fleet - AI-powered workflows
Proximal is a research lab for coding data. We build the data engine behind the next generation of autonomous coding agents.
Human intelligence for training and evaluating AI.
AI training data built from how experts reason, decide, and work.
Mechanize builds RL environments and sells them to leading AI labs.
Data lab to train frontier models and evaluate agents
Research and data infrastructure for training, evaluating, and advancing frontier AI models on long-horizon expert work.
Frontier AI that helps a billion weekly users think, create, code, and get work done.
Preference Model is building the next generation of training data to power the future of AI.
Bespoke Labs optimizes AI agents using cutting-edge research on data curation and evolutionary algorithms.
AI research and Claude models for coding and knowledge work, with safety at the core.
Automatically curate model training data to train faster, improve performance, and reduce costs.
Diffusing AI at scale across the services economy.
High-quality video, audio, image, and interaction data for frontier AI.
Environments and datasets for reinforcement learning.
Train, serve, and continuously improve AI models you own.
Aviro designs challenging worlds for AI models to explore and solve problems, delivered as post-training datasets and reinforcement learning
Steerable and explainable AI
A career network connecting people with opportunity in the AI economy.
Training data for recursive self-improvement
Push the frontier of your model
A modern data, compute, and AI agent platform for biotech R&D.
Real-world data, evaluation, and deployment infrastructure for physical AI.
Connecting billions through social platforms, with personal AI that can act for you.
NeoCognition is an AI agent lab for specialized intelligence.
Measuring how AI agents perform and behave in competitive, social environments.
We build the datasets, evaluation environments, and data pipelines that move AI agents from controlled demos to real-world production.
Prompt to silicon: agents that go from workload to GDS in weeks.
Train, evaluate, and deploy AI models you own.
An applied AI research lab building dependable, domain-grounded agents.
Research lab and product company building the platform for continual learning.
Vmax is automating reinforcement learning by transforming proprietary data and evaluations into new environments.
Robotics data and evaluation infrastructure for physical AI.
Chris Barber's picks38 companiesSource
