At least 5 years of combined product and engineering experience.
Formal computer science or software engineering training, or equivalent verifiable technical proficiency.
Familiarity with LLM and Agent technical principles, including LLM APIs, context windows and KV Cache, Agent loops, tool use, planning, memory, skills, MCP, subagents, and multi-agent collaboration.
Firsthand experience with prompt, context, or harness engineering, and an understanding of an Agent task's full computation path, including how model inference and tool execution alternate, where state resides, and where costs arise.
Understanding of distributed systems fundamentals, including state machines, idempotency, interruption recovery, and multi-tenant isolation.
Heavy usage of Agent products such as Claude Code, Cursor, or Manus as part of daily work, with the ability to explain tradeoffs in their harness designs.
Ability to create a runnable prototype using vibe coding.
Ability to make product decisions that trim features to user scenarios rather than accumulate platform capabilities.
Provide past system design documents or technical writing as a portfolio.
Responsibilities
Own the Agent execution-chain architecture, designing data, compute mechanisms, and corresponding tool architecture around the Agent loop so each task is completed correctly and economically.
Own the Agent runtime layer, architecting runtime approaches for business scenarios, including scheduled and event triggers, lifecycle state machines, heartbeat and interruption recovery mechanisms, and cloud-hosted environments.
Own Agent platform product architecture and experience for internal and external users to create, debug, publish, invoke, and iterate on Agents, including an Agent Store.
Build an open data ecosystem for the investment domain using OpenAPI, MCP, CLI, and skills, and design unified standards for open services.
Define Ground Truth and quality acceptance standards with the evaluation team, using real production consumers for each capability as the criterion for readiness.
Collaborate with the model team to tune models and harnesses.
Desired Qualifications
Experience transitioning from engineering to product, or experience with developer, platform, or infrastructure products.
GitHub, open-source project, or self-built Agent or MCP server work, and engagement with the open-source community.