Senior ML Engineer - Model Compression

About the Team 

The Compression and Parity team in GM’s Autonomous Vehicle (AV) Organization enables repeatable, high-velocity model deployments through principled and automated model compression under strict safety guarantees. We partner closely with model developers and deployment and infra engineers to ship numerically robust, low-latency models to the car, blending rigorous analysis with state-of-the-art methods and our own innovations. 

About the Role 

Over time, you will help grow and evolve the Compression and Parity function through the following: 

•  Developing and iterating on quantization and compression strategies for our AV models, considering model numerical properties, safety and latency constraints, and hardware performance, and partnering on deployment of quantized models to NVIDIA‑based AV hardware with our deployment, compiler, and kernel teams 

• Advancing our numerical sensitivity analyses to recommend safe compression policies per op/layer/block, using AV-relevant metrics (perception, trajectory, etc.) to evaluate compressed models, and collaborating with Embodied AI to support compression-aware modeling 

• Evolving sensitivity analysis, compression, and parity tooling into a connected, automated flow that makes low‑precision deployments repeatable, reliable, and low‑touch, with an emphasis on robust execution and maintainability 

• Bridging the gap between state-of-the-art model compression research and safety-constrained deployment while making strong technical contributions in cross-functional projects and educating others on best practices  

Your Skills & Abilities (Required Qualifications)     

•  Bachelor's degree in Computer Science, Electrical Engineering, Physics, Mathematics, Data Science / ML, or a closely related quantitative field (or equivalent experience)  

• 3+ years of industry experience focused on model optimization and deployment, with significant hands‑on work in neural network quantization / model compression / efficient inference  or relevant experience

• Strong proficiency in PyTorch and experience with graph‑level representations (e.g., PyTorch FX, ONNX) for capture and manipulation 

• Background in numerical linear algebra and optimization (conditioning, spectral properties, Jacobians, Hessians) and how they relate to quantization rob