[Operational] Superintelligence.

Backed by Y Combinator

Haladir is an AI product lab for verifiable domains, enabling RL and models to scale to economically-complex tasks via operations research and formal methods.

Research

• ConstraintBench: A Benchmark and RL Environment for Constrained Optimization from Formally Verified Rewards - A dual-purpose system for benchmarking and training LLMs on constrained optimization with solver-verified rewards.
• RLFR: Reinforcement Learning from Formally-Defined Rewards for Code Generation - Exploring a new paradigm for training coding models using formally verified reward signals.