AI Engineer

Harakumo; New York (NY); Remote; USD 140,000 - 230,000; Full time

Benefits offered by this job

Remote-first

Quarterly NYC meetups

Equity discussion in early talks

Job summary

Harakumo is seeking a strategic engineer to own and evolve its AI platform, including the AI Gateway, Vector DB, Memory, hosted Agents, and MCP connector. You will build features that let assistants run complete apps from conversations and improve tooling for safety and cost control.

We value hands-on experience with LLM APIs, retrieval systems, and agent frameworks, plus a bias for shipping real features used by customers.

Qualifications

• Production experience building with LLM APIs (Anthropic, OpenAI, Google or open models).
• Hands-on experience with retrieval (RAG), vector databases or agent frameworks.
• Bias for shipping: launched AI features used by real people.

Responsibilities

• Own Harakumo’s AI platform components (gateway, vector DB, memory, hosted agents).
• Grow the MCP connector and widgets to turn ideas into live apps.
• Ensure cost awareness and safe usage with spending controls and error handling.
• Write evals and conformance checks to prove model capabilities and maintain honest docs.

Skills

LLM APIs experience

RAG and vector DB

Shipping AI features

Tools

MCP servers

Job description

Own Harakumo’s AI platform: the AI Gateway, Vector DB, Memory, hosted Agents and the MCP connector that lets Claude, ChatGPT and other assistants build and run whole apps from a conversation.

• Build the AI services developers call through one key: an OpenAI-compatible gateway with routing, tool calling and structured output; vector search; memory for each end user; hosted agents
• Grow the MCP connector and its widgets so an assistant can take a one-line idea to a live, working app on Harakumo
• Make AI spend safe: holds, metering, spending limits and honest errors that never surprise a customer with a bill
• Write evals and conformance checks that prove what each model really supports, and keep the docs truthful about it

What you bring

• Production experience building with LLM APIs (Anthropic, OpenAI, Google or open models): tool calling, streaming, structured output, embeddings
• Hands-on experience with retrieval (RAG), vector databases or agent frameworks, and a sense of where they break
• A bias for shipping: you have launched AI features real people use, and you can show them
• Bonus: MCP servers, evals, cost and latency work on model traffic

What we offer

• Competitive salary and early-team equity (discussed in the first conversation)
• Remote-first, async-friendly; quarterly in-person in NYC
• Ownership of the AI platform from API to dashboard to docs