Product Manager

NGPartners; Palo Alto (CA); Hybrid; USD 180,000 - 280,000; Full time

Job summary

Ngpartners is seeking a Product Manager for the Agent Harness in robotics. You will own the platform's evolution, prototyping workflows, and coordinating with researchers and engineers to drive production releases of autonomous agent capabilities.

The ideal candidate has deep experience with agent-like systems and a track record of shipping complex features, while collaborating across multi-disciplinary teams focused on hardware, simulation, and software stacks.

Qualifications

• Prior experience building Codex, Claude Code, Cursor, Devin, Lovable, or a comparable agent system with direct ownership of shipped production features.
• Deep comprehension of agent systems and trade-offs in planning, tool selection, context/state management, execution sandboxes, verification, and recovery loops.
• Hands-on ability to prototype agent workflows, integrate APIs and tools, and diagnose execution trace failures.
• Experience improving reliability via evaluation suites, telemetry metrics, and user feedback loops.
• Strong product design sensibilities for autonomous systems—mastering delegation, inspectability, intervention boundaries, and error recovery.
• Ability to absorb unfamiliar technical domains quickly through research literature, code, and collaboration.
• Proven track record partnering with technical teams to deliver milestone releases.

Responsibilities

• Own the agent harness end-to-end, guiding it from architectural vision to releases with clear acceptance criteria across research and engineering teams.
• Define how the agent plans, uses tools, manages context, executes commands, verifies outputs, and recovers during long-running tasks.
• Shape the developer experience for autonomous interaction, allowing engineers to delegate complex real-world work and correct errors.
• Identify high-impact use cases across engineering research, robot design, simulation training, and deployment.
• Improve system reliability with automated evals, feedback loops, and telemetry to boost task completion and reduce latency.
• Prototype agent workflows, integrate specialized tools, and diagnose failure modes through code and traces.
• Rapidly learn robotics concepts to collaborate with Brain, Hardware, and Platform specialists.

Skills

Agent Track Record

Systems Intuition

Telemetry Leadership

Autonomous Workflow

Eval & Telemetry Leadership

Milestone Releases

Technical Absorption

Job description

Product Manager, Agent Harness (Robotics)

EMPLOYMENT TYPE: Full-Time

WORKPLACE SETUP: Hybrid

DEPARTMENT: Platform

About Our Client

Our client is an Applied R&D lab building Asimov, an open-source humanoid robot platform, alongside the complete software stack that powers it. Their foundational mission is to make humanoid labor economically viable, turning software into physical labor at global scale. They develop across the full technology stack: hardware architecture, locomotion, autonomy, simulation, and high-performance infrastructure. They move fast, ship iterations directly to real physical robots, and open-source their technology.

The Role

Coding agents fundamentally changed how software gets authored. Our client is building the equivalent for physical robotics. Their agent harness for robotics is an agent platform that investigates engineering solutions, assists in designing manufacturable robots, trains them inside physics simulations to perform physical manipulation, and coordinates deployment into real-world operating environments.

As Product Manager for the Agent Harness, you will take full ownership of this platform's evolution. Our client is seeking a deeply technical PM who has built systems like Codex, Claude Code, Cursor, Devin, or comparable coding agents, and who can rapidly transfer that expertise into robotics and physical AI. You will partner with technical users, researchers, and engineers, prototype workflows firsthand, and drive ambitious concepts into production releases.

Key Responsibilities

• Product Ownership: Own the agent harness end-to-end, steering it from an ambitious architectural vision into focused releases with measurable acceptance criteria across research and engineering teams.
• Core Agent Mechanics: Define how the agent plans, utilizes tools, manages context window state, executes commands, verifies outputs, and recovers during long-running engineering and physical tasks.
• Autonomous Interaction & DX: Shape how engineers delegate complex real-world work, inspect intermediate outputs, intervene, and correct errors, delivering an intuitive developer experience.
• Value Discovery: Work directly with technical users to pinpoint initial high-impact use cases across engineering research, robot design, simulation training, and deployment.
• System Reliability: Drive agent dependability through automated evals, user feedback, and runtime telemetry—raising task completion rates, lowering human intervention, and optimizing latency and compute.
• Hands-On Prototyping: Prototype agent workflows, integrate specialized tools, and independently diagnose failure modes directly via code and execution traces.
• Cross-Stack Fluency: Rapidly master robotics concepts—from VLAs and world-action models to simulation and manufacturability—collaborating with Brain, Hardware, and Platform specialists.

Candidate Profile & Qualifications

• Coding Agent Track Record: Prior experience building Codex, Claude Code, Cursor, Devin, Lovable, or a comparable agent system as a PM, software engineer, or technical founder, with direct ownership of shipped production features.
• Deep Systems Intuition: Thorough comprehension of agent systems, capable of articulating technical trade-offs across planning, tool selection, context and state management, execution sandboxes, verification, and automated recovery loops during long-running tasks.
• Hands-On Technical Aptitude: Comfortable navigating complex unfamiliar codebases, building agent prototypes, integrating APIs and tools, and diagnosing execution trace failures.
• Telemetry & Evals Leadership: Proven experience raising agent reliability through comprehensive evaluation suites, telemetry metrics, and user feedback loops to improve task completion and manage resource costs.
• Autonomous Workflow Judgment: Strong product design sensibilities for autonomous systems—mastering user delegation, inspectability, intervention boundaries, and error recovery.
• Accelerated Technical Absorption: Demonstrated capacity to ingest unfamiliar technical domains swiftly via research literature, code repositories, and collaborative engineering discussions.
• Outcome-Driven Leadership: Strong record of partnering with technical teams to convert ambitious technical ideas into iterative, measurable milestone releases.

Preferred Background (Nice to Have)

• Physical AI & Robotics: Familiarity with inverse kinematics, motion planning, control theory, reinforcement learning, imitation learning, vision-language-action models (VLAs), or world-action models (WAMs).
• 3D Perception & Scene Understanding: Experience with room scanning, 3D reconstruction, or transforming physical spaces into synthetic simulation assets.
• Simulation Infrastructure: Practical background in simulation environments, training pipelines, reproducible experiments, skill composition, and hardware validation.
• Hardware Manufacturability: Exposure to robot design constraints, CAD tooling, component selection, bills of materials (BOMs), assembly, and hardware sourcing.

WHY JOIN OUR CLIENT

You will establish what an autonomous agent platform becomes when its terminal output is a functioning humanoid robot. The harness touches every layer of how the robot gets developed: research, mechanical design, simulation training, and physical deployment. The category is brand new, the team deploys directly to physical hardware every week, and the technology is built in the open.

Approach to AI Collaboration

Deep AI expertise is not required across all functions, but our client expects every team member to be intellectually curious, drawn to experimentation, and excited to leverage AI as a day-to-day thinking collaborator. For this role, AI fluency is fundamental. Team members who succeed do not treat AI as a mere novelty; they employ it to sharpen reasoning and build systems that enable others to scale.

Equal Opportunity & Accommodations

Our client hires exceptional people from diverse backgrounds. Candidates excited by this mission are encouraged to apply even if they do not meet every single bullet. Our client is an equal