Sciloop

Site unreachableAI-native

Expert STEM reasoning data for frontier AI labs.

What it does

Sciloop creates expert-level math and physics problems that frontier AI models can't solve, then sells the data to AI labs for training and evaluation. Our problems are created by IPhO and IMO medalists — the top 0.01% of STEM talent globally. On our benchmark, models like GPT 5.4 Pro and Gemini 3.1 Pro score 0-5% on our hardest problems. We work with AI labs to supply continuous, fresh training data that pushes the frontier of mathematical and scientific reasoning. Founded by Bilal and Osman, International Physics Olympiad medalists from MIT with hands-on ML research experience at MIT CSAIL.

Next to it in Data for AI · Expert STEM synthetic data for AI training

  • Liva AIY Combinator S25

    Audio & Video Data

  • SenseiY Combinator S24

    Robotic Training Data at Scale

  • PerfectBit, Inc.Y Combinator X26

    Correct by construction AI training data

  • Cortex AIY Combinator F25

    Large-scale real-world robot & human data for embodied AI

  • BesampleTechstars TS 2024

    Besample | Research beyond the West

  • Luvua16z speedrun SR005

    The Human Data behind Frontier AI

  • LuelY Combinator W26

    Turning everyday words and actions into usable training data.

  • Vision LabY Combinator X25

    Industrial data layer for robotics training