Robotics Software Engineer, Mission Evaluation & Fine-Tuning

Description

The Irvine team works on embodied AI that runs on real robots and sensors, and the systems are deployed in the field.

The company is based in Southern California's robotics hub and builds risk-aware, dependable AI systems for hard robotics problems.

Your work ships, gets tested on hardware, and improves from real-world runs.

The company pairs careful engineering with learning methods already used in deployed products worldwide, rather than relying only on data-driven or transformer-only designs.

The company builds foundation models for perception, planning, localization, and manipulation, with a focus on explainability and safety.

The team spans AI, software, robotics engineering, product, field deployment, and technical communication.

Responsibilities

• You own the success metric for end-to-end inspection behavior.
• You build the evaluation loop: golden-labeled datasets, seeded-fault runs, and regression suites at feature, system, and release levels, gating every release candidate.
• You turn field evidence into training and evaluation sets.
• You connect technical performance to the metrics the customer measures.
• You make failures clear by reviewing mission logs, assigning each miss to navigation, localization, perception, or behavior, and sending it to the owning team with evidence.

Requirements

• You hold an MS in computer science, robotics, aerospace engineering, or a closely related field, with 3+ years in robotics or ML engineering.
• You write strong Python and working C++, and you are comfortable with ROS 2 logs, bags, and replay tooling.
• You have evaluated and fine-tuned vision or vision-language models using PyTorch, LoRA or similar, and built the datasets and metrics behind them.
• You understand statistics well enough to tell a benchmark number from a field result and to design experiments that separate them.
• You build data pipelines that move evidence from robot to training set to dashboard.
• Set you apart: you have authored or maintained a robotics benchmark, evaluation suite, or challenge.
• Set you apart: you have done simulation-based evaluation or sim-to-real validation.
• Set you apart: you have worked on MLOps for on-edge models, including quantization, versioning, and fleet rollout.
• Set you apart: you have industrial inspection, gauge reading, or anomaly detection experience.

Conditions

The salary range is generous, and final pay depends on your background and experience.

Base pay may also vary with role scope, job-related knowledge, skills, experience, and the Irvine, California market.

The headquarters is in Irvine, California, and the team works closely with colleagues across the US and abroad.

The company evaluates people on merit and does not discriminate on any legally protected basis.