SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 125,000 - 250,000 / annual
CircuitHub is reshaping electronics manufacturing with The Grid, a factory-scale robotics platform that dramatically improves the efficiency of small-batch, high-mix electronics assembly. The company has raised $45M from top-tier investors including Plural, Y Combinator and Google Ventures, is already profitable, and counts major EV makers, top AI labs, and Fortune 500 companies among its customers.
In this role, you will build AI agents that automate real manufacturing and business processes. These agents are production systems with concrete jobs: inspecting circuit boards via computer vision, programming robots by reading datasheets, monitoring factory operations, managing back-office tasks like invoice processing and supplier coordination, and even writing code. You will find manual processes, automate them end-to-end, and continuously improve them for cost, speed and accuracy.
Your approach will emphasize rigorous evaluation: agents earn the right to work autonomously through blind testing against human verdicts, shadow runs, and replay analysis. You'll build evals before trusting results, decide what requires AI judgment versus plain code, and ensure safety through sandboxed cloud environments and idempotent tool calls. You'll also improve the platform that all agents run on and work closely with factory operators, finance teams and other stakeholders.
Recent examples of agent work include: image alignment that identifies the worst 20% of parts (which hold 90% of defects) and caught all 13 defects operators had marked in blind testing; photo checks for stray parts in robot storage scoring 10/10 blind at ~8¢ per photo; invoice extraction from supplier emails with 66/66 accuracy; and a constraint solver that replans the entire factory every two minutes.
The tech stack includes Claude, OpenAI APIs, MCP, TypeScript (Bun, React), Python (FastAPI, Dagster, OpenCV, OR-Tools), PostgreSQL, Nix/NixOS, Nomad, Terraform, Vault, AWS, Modal sandboxes, and OpenTelemetry.
REQUIREMENTS:
- Proven experience putting LLM agents into production; ability to diagnose what breaks and validate that systems work
- Strong software engineering fundamentals (concurrency, idempotency, data modeling, security) — this role is primarily real engineering, not prompting
- Discipline in building evals before trusting results
- Interest in the physical world (cameras, robots, mechanical constraints)
- Bonus: computer vision, optimization, or electronics domain knowledge
