Agentic Engineer
Job Description
As an Agentic Engineer at Zeromatter, you'll build the agents, harnesses and evals that make Zeromatter an AI-native company, and the agentic tools customers use with our simulation product. You'll work across three areas: our internal agent platform and AI software factory; the infrastructure and models agents run on; and agentic tools and interfaces for our simulation product, such as building and debugging simulations, generating data with simulation, and Model Context Protocol (MCP) servers. You'll evaluate new options on real workloads and shape what we build on.
Responsibilities
Build multi-agent systems for our AI software factory, where agents plan, write, test and review code, and for a shared company knowledge base.
Design the multi-agent harness: skills, orchestration, memory, and coordination between long-running, self-learning agents that improve from feedback and real use.
Build sandboxes and secrets management that let agents run safely, quickly and at scale.
Evaluate the latest models, coding agents and agentic platforms, and recommend what we adopt and how we route between them on quality and cost.
Build domain agents across engineering and operations, from automated validation and code review to customer support.
Build agentic tools and MCP servers for our simulation product.
Own the evals that show each agent works, and turn failures from real use into regression tests.
Qualifications
Ideally, you have an engineering degree in a related field (e.g. computer science, mechanical engineering, physics, math, robotics, etc.)
Senior or staff-level individual contributor experience, owning systems end to end.
A track record of shipping AI agents to production that people rely on every day.
Rigor in evaluation: you measure before you claim a result.
Experience writing and deploying services, containers or Kubernetes manifests.
Proficiency in Python, with TypeScript or Rust a plus.
Experience with sandboxes or isolation (gVisor, Kata, microVMs).
Experience with LLM gateways, model routing or serving.
Experience with vision-language or computer vision models, simulation or 3D.
