Responsibilities
- Identify the biggest user pain points where a crypto AI can materially improve outcomes.
- Turn ambiguous ideas into a clear MVP, with crisp scope, constraints, and success metrics.
- Prototype full AI experiences in Python to validate value and quality before we ship to production.
- Own prompts and context engineering: instruction design, context shaping, guardrails, tool/function calling patterns, and output formatting.
- Build practical evaluation loops: golden sets, scenario coverage, qualitative rubrics, regressions, and acceptance criteria.
- Design the AI user experience: make it clear, trustworthy, and resilient if things go wrong.
- Run fast experiments, learn from real outputs and usage data, and iterate quickly.
- Partner with Engineering to ship: provide handoff specs, edge cases, evaluation results, and support debugging and iteration post-launch.
- Work on whatever surface is the highest leverage.
Requirements
- Strong product judgment and the ability to make good calls under ambiguity.
- Hands-on Python prototyping ability: you move fast, write clean code, and can translate ideas into working prototypes.
- Practical LLM experience + intuition: you understand prompt iteration, context design, and have a strong intuition for how to build useful products on top of LLMs.
- A strong evaluation mindset: you can define quality, test for failure modes, and prevent regressions without heavy process.
- High-agency execution: you can go from 2vague problem 2 1shipped learning 2 with minimal supervision.
- Excellent communication skills (verbal and written): convey complex messages clearly and simply, and driving conviction across stakeholders.
Nice to have
- Shipped user-facing AI features (chat, agents, copilots, summarization, search/Q&A, personalization).
- 0 to 1 experience in fast-moving environments and owning ambiguous problems end-to-end.
- Experience building tool-using and agent-like workflows.
- Experience and interest in cryptocurrency.
