Artificial Intelligence
We post-train and align large models, build the algorithms around them, and embed both into production — where a model earns its keep or is discarded. The same stack serves clients and runs our own capital and on-chain systems: research here is not a paper. It is the product.
- Post-training · SFT · RLHF · DPO
- Reasoning & alignment
- Autonomous agents
- Inference optimization & serving