Product Manager: AI Agents & MCP Tooling

Recorded Future; Boston, Northern (MA, KY); Hybrid; USD 129,000 - 194,000; Full time

Job summary

Recorded Future is seeking a hands-on Product Manager to own the lifecycle of AI agents and MCP tools, applying large language models to real threat intelligence problems. You’ll build and evaluate agents, refine tool descriptions, and ensure the MCP surface meets customer needs.

This mid-level role demands comfort with GitHub and prompt engineering, strong product judgment, and a bias for shipping value quickly. Expect to collaborate across teams and measure impact.

Qualifications

• Hands-on experience building LLM-powered agents or workflows.
• Working knowledge of major LLMs and cost tradeoffs.
• Ability to write clear tool descriptions and assess coverage gaps.
• Comfort reading code and collaborating with engineers on design and evals.

Responsibilities

• Own end-to-end lifecycle of AI agents: design, build, evaluate, validate, deploy.
• Build agents that orchestrate LLMs and tools for real intelligence use cases.
• Refine MCP tool surface and tool descriptions; improve coverage.
• Analyze usage patterns to identify gaps and needs for new tools.
• Write and expand evaluation suites; guide iteration and deployment.
• Test with customers and ensure outputs meet workflows and expectations.
• Collaborate in GitHub with engineers on agent logic and tool definitions.
• Define and track performance metrics like accuracy and latency.
• Prioritize roadmap to deliver maximum customer value.

Skills

Agent Builder

Model-Literate

Tool-Design Sense

Technically Comfortable

Evaluation-Minded

Data-Informed

Customer-Oriented

Strong PM Skills

Pragmatic and Outcome-Driven

Cybersecurity experience

Tools

GitHub

Prompt Engineering

MCP (Model Context Protocol)

Job description

Recorded Future is seeking a hands-on Product Manager to own the lifecycle of AI agents and MCP tools, applying large language models to real threat intelligence problems. You’ll build and evaluate agents, refine tool descriptions, and ensure the MCP surface meets customer needs.

This mid-level role demands comfort with GitHub and prompt engineering, strong product judgment, and a bias for shipping value quickly. Expect to collaborate across teams and measure impact.