You will shape how Kinro's systems think, reason, and improve. The research question underneath everything we do: how far can a machine be trusted to run a regulated business, and how do you prove it? You will own the full loop from hypothesis to measurement to production, working on grounded generation, compliance-aware reasoning, and multi-turn conversations about commercial insurance. This is not a role for researchers who want to stay in the lab.
What you'll do
- Design and run evaluations that measure what actually matters: bound policies, compliance, and customer satisfaction.
- Research and implement improvements to our core capabilities.
- Ensure every model update is measurable and safe to ship.
- Stay current with the frontier of LLM research and integrate relevant advances quickly.
- Collaborate with engineering to take research from notebook to production.
What we're looking for
- PhD or equivalent research experience in ML, NLP, or a related field.
- A track record of shipping research into real systems. Papers are great, deployed models are better.
- Deep familiarity with LLMs, fine-tuning, and evaluation methodology.
- Strong Python skills; comfortable with PyTorch or JAX.
- Clear thinking and clear writing. You can explain your ideas to engineers and customers alike.
Nice to have
- Experience with agentic systems or tool use.
- Knowledge of regulatory constraints in insurance or financial services.
- Prior work at a research lab (DeepMind, Meta AI, OpenAI, etc.).