We build post-training data. Got a domain you can't staff for? We'll find the experts. Got an eval you can't move? We'll build the failure cases against it. Need a custom annotation tool? We'll build it.
Our engineers sit with your team to figure out where the model is breaking. Every batch is run against your eval before delivery. The experts are yours, not subcontracted.
If you're looking for a high quality dataset, contact us.
Founder & CEO
xAI - Manus AI - Enhanced Radar YC W25
Former xAI iOS engineer who contributed to Grok Imagine and Grok 4.1. Core agent engineer at Manus AI. Head of iOS and AI training work at Enhanced Radar, where he built its first aviation world model.
Co-founder
rct.AI YC W19 - Boom Foundation BNB MVB S8
Built across AI, gaming, incentive systems, and enterprise BD. Led product, gameplay loops, data collection, and BD at Boom Foundation. Later worked on a B2C AI health app focused on reliable consumer AI.
Help train frontier AI, build evaluator skill, and share in the upside your judgment helps create.
The first Alphaset Fellows will start with code, reasoning, STEM, and safety evaluation tasks. Qualification comes through short sample evals, but the waitlist is open to students, researchers, domain experts, and people leaving big tech who want recurring AI income.
This is not a generic labeling pool. We are building a verified contributor network for high-quality AI evaluation.
