Staff ML/AI Software Engineer Autonomy Evaluation Platform

Staff ML/AI Software Engineer

As a Staff ML/AI Software Engineer within the Evaluation team under the Simulation, Evaluation, and Data organization, you will shape the design and delivery of the evaluation strategy for GM's autonomous vehicle programs. In this role you'll be responsible for directing software validation and model training, implementing data mining strategies and building metrics that measure autonomy performance in simulation and real-world environments. You will partner with ML engineers, systems engineers, and research scientists to build reliable, reproducible, and high-signal evaluation mechanisms that accelerate model iteration and improve safety, performance, and reliability across GM vehicles.

The Evaluation team is dedicated to creating, maintaining, and evolving the evaluation ecosystem that underpins GM's pursuit of safe, high-performing, and scalable driverless technology. The team delivers trusted metrics, automated workflows, and scalable tools that enable data-driven decision-making at every stage of AV development. Evaluation team members collaborate closely with Simulation, Motion, Perception, and Release teams, acting as system-level integrators and arbiters of end-to-end AV system quality. The team treats road, data mining, training, and metrics as equal use cases for our analytics framework and evaluation goals. By joining this team, you will guide the evolution of core evaluation platforms and frameworks, champion the interpretation and communication of system-level results, and play a central role in accelerating GM's progress toward safe, validated AV deployment at scale.

Act as technical architect, defining technical vision and strategy and aligning the team with broader company objectives.

Design and implement scalable, reliable data pipelines and indexing/aggregation services to support model training and evaluation at scale, with strong guarantees for data quality, lineage, and reproducibility.

Leverage vision-language models (VLMs) and large language models (LLMs) to classify autonomy performance, mine critical scenarios, and prioritize validation efforts, integrating human-in-the-loop where appropriate.

Define and operationalize metrics and acceptance gates that quantify autonomy performance in simulation and on-road, integrate into CI/CD to guide release and merge decisions.

Build and maintain evaluation dashboards and reports that provide clear, explainable insights to engineering and leadership, including trend analysis, drift detection, and scenario coverage.

Maintain a high technical standard through architectural design, design reviews, and code reviews, setting patterns and best practices for the broader team.

Collaborate cross-functionally to centralize investments that serve multiple stakeholders, align roadmaps, and reduce duplicated efforts.

7+ years professional experience developing Python and C++ in production environments, including unit testing, code reviews, performance tradeoffs, and reliability practices.

5+ years applied experience in data analysis, ML evaluation, or autonomy analytics, working with large-scale datasets and statistical methods, leveraging tools like SQL.

Demonstrated technical leadership delivering measurable impact across teams, including setting standards and influencing architecture.

Strong written and verbal communication, driving decisions, communicating risk, and giving constructive feedback to diverse stakeholders.

Bachelor's or higher degree in Computer Science, Data Science, Mechanical or Aerospace Engineering, or equivalent practical experience.

Experience with computational geometry, linear algebra, PyTorch, and machine learning for perception, prediction, or planning.

Applied data analysis in robotics or electro-mechanical systems, including sensor data (camera, lidar, radar) and time-series analysis.

Expertise in profiling, analysis, debugging, and performance optimization across Python/C++ and distributed workloads.

Familiarity with ROS, Pandas, NumPy, SciPy, Python bindings for C++, and plotting/visualization libraries.

Experience with scenario mining, evaluation metric design, and release gating for autonomy systems.