Lead AI Engineer

Onto Innovation•Hillsboro, OR

3d•Onsite

About The Position

Onto Innovation is hiring a hands-on technical leader to define and build AI systems for semiconductor metrology. The role works closely with system engineers, algorithm developers, and applications teams to understand metrology workflows and data flows, then sets the vision and architecture for AI in metrology. The engineer designs, builds, evaluates, and productionizes AI systems—including agentic systems—to accelerate setup and troubleshooting and improve measurement quality. Agents should be grounded in real tools and data, with appropriate human oversight. The role encompasses AI strategy and architecture, production software development, agent workflows across the metrology lifecycle, integration with simulators, solvers, and tool data, and retrieval from fab knowledge sources with answer provenance. Systems must be designed for trust through human checkpoints, uncertainty, audit trails, and refusal when evidence is insufficient. Deployment may be on-premises or air-gapped, with models selected according to latency, cost, and data-control constraints and releases that accommodate customer change control. The Lead AI Engineer will build rigorous evaluation processes and maintain sound measurement science, including surrogate models, parameter inference, calibration, and calibrated uncertainty. The role sets technical standards, makes build-versus-buy decisions, conducts reviews, mentors engineers, and improves software quality and team processes. It also establishes responsible AI-assisted development practices and translates customer setup or yield problems into agent workflows whose results can be assessed against reference techniques. The position is on-site and relies on close, daily collaboration with applications and system engineering. It includes regular office time and occasional visits to a fab or customer site. Work is reviewed, documented, and qualified before release; shipped systems run in production lines measuring wafers around the clock.

Requirements

MS or PhD in computer science, electrical engineering, physics, applied mathematics, or equivalent hands-on experience.
At least 6 years of applied ML/AI experience in production, including at least 3 years in technical leadership, leading projects or owning technical direction of a significant AI system.
Experience defining AI strategy and system architecture and getting a technical vision adopted across engineering.
Hands-on experience with LLM-based systems, including agents, tool calling, retrieval, and prompt and model evaluation, shipped to real users and monitored after release.
Strong ML foundations, including deep learning in PyTorch, regression, anomaly detection, and calibrated uncertainty.
Experience with MLOps, including reproducible training, dataset and model versioning, model registries, CI/CD, monitoring, and GPU-aware inference optimization, and experience with agile and DevOps ways of working.
Track record of rigorous AI evaluation, including building test sets, defining metrics, and reporting where systems fail.
Deep Python skills and strong software engineering, including data structures, design patterns, and object-oriented design; ability to write clean, tested production code.
Hands-on experience with AI-assisted development tools, including prompting, critical review of generated code, and converting it into maintainable production code.
Working understanding of software security, intellectual property, data privacy, and responsible AI in enterprise software.
Strong analytical and troubleshooting skills for complex software systems and ability to independently research and solve difficult technical problems.
Experience developing software on Windows with Visual Studio and Azure DevOps.
Clear technical writing and ability to explain AI system decisions to process engineers.

Nice To Haves

Strong programming skills in C++, C#, or CUDA where models interface with instruments.
Semiconductor process control experience, such as metrology, inspection, SPC, yield analytics, or time spent in a fab.
Experience with scatterometry or OCD, ellipsometry, reflectometry, interferometry, or e-beam imaging.
Experience with tool-calling and MCP frameworks, multi-agent orchestration, vector and hybrid search, LLM fine-tuning, or on-premises LLM serving.
Experience with AI security and IP protection, regulated release processes, or customer data that cannot leave the site.
Experience with physics-informed ML, surrogate modeling, physics-based modeling and numerical simulation, or Bayesian inverse methods such as MCMC or variational inference.
Experience with parallel or high-performance computing.
Experience integrating AI into software development, such as AI-assisted code review, automated test creation, and requirements analysis.

Responsibilities

Define the AI strategy, reference architecture, and roadmap for metrology, including build-versus-buy decisions tied to customer problems.
Write, test, and ship production code for prototypes, agent workflows, evaluation tooling, and supporting systems.
Design agent workflows for guided recipe and model setup, qualification runs, tool matching, and root-cause triage when measurements drift.
Connect agents to simulators, solvers, and tool data, and ground them in manuals, recipes, service tickets, and historical measurement outcomes with provenance.
Design trustworthy systems with human-in-the-loop checkpoints, confidence and uncertainty, audit trails, and refusal when evidence is insufficient.
Deploy AI systems on-premises or in air-gapped environments, selecting local or hosted models according to latency, cost, and data-control constraints and supporting customer change control.
Build evaluation harnesses with golden task sets, tool-call correctness measures, comparisons of agent-built and expert-built recipes against TMU budgets, and regression tests for model or prompt changes.
Maintain sound measurement science through surrogate models, parameter inference, tool calibration, and calibrated uncertainty; own relevant ML, including domain-shift concerns across tools, layers, and fabs.
Set technical standards, conduct design and code reviews, mentor mid-level and junior engineers, and make technical approach and build-versus-buy decisions.
Improve software quality through unit testing, CI/CD test automation, legacy-code refactoring, and static and dynamic analysis.
Establish responsible practices for AI-assisted coding, debugging, documentation, and test generation, including human review, automated testing, static analysis, security scanning, and requirements validation.
Identify gaps in team processes and toolchains and drive improvements.
Translate customer setup or yield problems into tractable agent workflows and explain results to engineers evaluating them against reference techniques.
Own production AI systems after launch by monitoring them, diagnosing failures, and shipping updates.