Member of Technical Staff - AI Agent Systems

Description

About Warp

At Warp Sciences, our mission is to help the world discover new medicines faster—and build the capacity to do it here in America. We bring AI, robotics, and life sciences together in automated labs that design, run, and learn from experiments, with each experiment making the next one better. We’re not developing drugs of our own. We’re building the infrastructure for the people who do.

We're looking for a Member of Technical Staff, AI Agent Systems, to build and advance the fleet of AI agents that helps us develop, maintain, and improve our software and robotics systems. You'll build the agents, harnesses, and coordination infrastructure that form our software factory—the foundation for a self-reinforcing loop in which better software enables better experiments, and what we learn from those experiments helps improve the systems behind them.

This role is for an AI-native builder and systems thinker: someone who works fluently with frontier AI tools, understands how large software systems fit together, and can turn emerging agent capabilities into reliable engineering workflows across the cloud and the edge.

Responsibilities

• Develop, customize, and operate agents and agent harnesses that can take on substantial software engineering work, from implementation and testing to debugging, maintenance, and integration.
• Build the infrastructure for coordinating teams of agents, including task decomposition, shared context, tool access, communication, review, and handoffs.
• Develop and empirically evaluate communication and coordination policies—determine when agents should work independently, share information, ask for help, or escalate to a human.
• Connect agent workflows to our development infrastructure, including repositories, CI, simulation, deployment systems, telemetry, and robotics tools.
• Work across the product stack, from cloud services and data infrastructure to edge software and robot execution, helping integrate critical components into a coherent system.
• Close the loop between agent execution, evaluation, and improvement—understand failures, identify the right intervention, and make the software factory increasingly capable and reliable.

Requirements

• We care much more about what you've built than your title or number of years of experience.
• Strong software engineering foundations and experience building, debugging, and maintaining substantial software systems.
• An AI-native approach to engineering with deep expertise using frontier AI tools and orchestrating swarms of agents to tackle difficult agent-driven software development and production tasks.
• Experience building or customizing agents, agent harnesses, communications protocols, or other systems that combine models, tools, context, and execution.
• Strong systems thinking—the ability to understand dependencies across components, identify where interfaces break down, and make sound architectural decisions.
• Strong experimental instincts and the ability to evaluate whether an agent, workflow, or coordination strategy actually improves outcomes.
• Experience working closely with other engineers to ship and operate a shared software product.
• Good judgment about verification, permissions, deployment, and when a system needs human oversight.
• High ownership and excitement about problems where the right approach isn't obvious yet.
• Helpful, but not required
• Experience in any of the following is a plus:
• Agent evaluation, execution tracing, failure analysis, and reproducible benchmarks
• Context management, memory, retrieval, and tool design for long-running agents
• Large software products, distributed systems, and complex codebases
• Developer infrastructure, CI/CD, automated testing, and deployment
• Cloud-to-edge systems, robotics software, simulation, or hardware integration
• Sandboxed execution, permissions, and recovery for autonomous workflows
• Human-agent collaboration and workflows for reviewing agent-generated work
• Improving agent behavior through data, feedback, or model adaptation
• Example Problems You Might Work On
• How do we turn a high-level engineering objective into work that a team of agents can execute, verify, and integrate?
• When should agents communicate, what should they share, and how do we measure whether that communication helps?
• How do we give agents enough context to work across a large codebase while keeping that context accurate as the software changes?
• How do we coordinate a change that spans cloud services, edge software, and robot behavior—and verify that the resulting system works?
• How do we distinguish a plausible implementation from one that is correct, maintainable, and ready to deploy?
• How do we use failures from software development and real-world robot operation to improve the agents, tools, and workflows that support the next iteration?