ML/AI Engineer at a seed-stage team working on one of the hardest problems in crypto - backed by tier 1 VCs.

You are the person who takes ML research and turns it into production systems that work. Training pipelines, eval harnesses, inference services, synthetic data generation, feedback loops. Not a research role. Not a backend integration role. The full path from prototype to deployed system.

Needed:

-> proven ability to ship ML models into real products, not just notebooks and demos

-> reinforcement learning experience: applied RL, DPO, GRPO, reward modeling, preference optimization

-> LLM post-training workflows: SFT, preference datasets, rejection sampling, multi-candidate eval

-> you can read a paper, decide if it's useful, build the first version, and keep it reliable as requirements change

Crypto or DeFi familiarity a strong plus. PST/EST timezone overlap required.