Staff Technical Product Manager, AI Agent

Hercules; United States; Hybrid; USD 100,000 - 250,000; Full time

Benefits offered by this job

AI token budget

Healthcare, vision, and dental

Daily in-office meals

Equity

Job summary

Hercules in the United States seeks a senior product leader to advance an AI coding agent. You will own the context, prompts, memory, tooling, and evaluation systems to improve cost, speed, and accuracy across production and offline tests.

Lead end-to-end system design with engineering, guided by data from millions of interactions and repositories to prioritize experiments and shape the product roadmap. This is a hands-on, senior role.

Qualifications

• Strong software fundamentals with system design, data modeling, and APIs.
• Sharp product sense and a history of shipping simple, delightful user experiences.
• Hands-on experience with context engineering: summarization, memory, and tooling.
• Familiar with agent frameworks, evaluation platforms, and large-scale data usage.

Responsibilities

• Drive continuous improvements to cost, latency, and accuracy of the AI coding agent.
• Own end-to-end design of the context and harness system, including prompts, tooling, memory, and integration.
• Design and operate online and offline evaluation systems for production-quality agents.
• Leverage data from millions of conversations and repositories to guide experiments.
• Collaborate with engineering on complex implementation and architecture decisions.

Skills

Software fundamentals

Product sense

Context engineering

TypeScript

React

Agent frameworks

Evaluation platforms

Tools

TypeScript

React

Job description

Role overview

A senior product role focused entirely on advancing an AI coding agent, with the primary objective of improving its Pareto frontier of cost, speed, and accuracy. The position owns the full context and harness system design, including prompts, tools, memory, skills, summarization, and hooks, and is responsible for both online and offline evaluation systems. The work is data-rich, drawing on a large user base, hundreds of thousands of code repositories, and millions of chat threads to guide experimentation.

Responsibilities

• Drive continuous improvements to cost, latency, and accuracy across the AI coding agent.
• Own the end-to-end design of the context and harness system, including prompt strategy, tooling, memory, and codebase integration.
• Design and operate online and offline evaluation systems that measure agent quality in production.
• Use large-scale user, repository, and conversation data to prioritize and validate experiments.
• Partner closely with engineering counterparts on complex implementation and architecture decisions.

Requirements

• Strong software fundamentals with the ability to code and reason about system design, architecture, data modeling, and APIs.
• Sharp product sense and a track record of shipping simple, delightful user experiences.
• Hands-on experience with context engineering practices such as summarization, prompt caching, subagents, memory management, and tool design.
• Working knowledge of TypeScript and React, plus familiarity with agent frameworks and evaluation platforms.
• Active personal use of frontier AI coding tools to multiply personal development output.

Nice to have

• Experience operating at the pace of a fast-moving AI product area where best practices evolve weekly.

Benefits and work setup

• San Francisco based with a strong in-office preference; remote considered for exceptional candidates willing to work US time zones and travel to the office regularly.
• Unlimited AI token budget, flexible vacation policy, healthcare, vision, and dental coverage, plus daily in-office meals.
• Salary range of $100,000–$250,000 plus equity, varying with seniority and location.