Senior/Staff Software Engineer (Machine Learning Runtime), Motion Planning

Responsibilities

• Design, implement, and optimize the production runtime pipeline for ML-based autonomous vehicle planning.
• Integrate machine learning models into a scalable, safety-critical planning software stack.
• Optimize end-to-end inference performance through algorithm optimization, C++ implementation, profiling, memory optimization, and efficient execution on compute-constrained embedded platforms.
• Develop high-performance C++ components for feature extraction, inference, post-processing, trajectory generation, and runtime orchestration.
• Design runtime validation, safety guardrails, and fallback mechanisms to ensure generated trajectories are feasible, robust, and compliant with traffic rules.
• Investigate complex runtime issues by tracing failures across the planning pipeline, from feature extraction through inference to trajectory generation.
• Debug difficult edge-case scenarios using logs, profiling tools, visualization, offline replay, and quantitative analysis to identify root causes and implement robust solutions.
• Build tooling and infrastructure to improve debugging, observability, performance analysis, testing, and runtime monitoring.
• Collaborate closely with machine learning engineers to efficiently deploy new models into production while maintaining system reliability and performance.
• Drive continuous improvements in runtime architecture, scalability, reliability, and maintainability.
• Ensure technical work complies with the company's Quality Management System (QMS), customer requirements, regulatory standards, and internal engineering processes.

Requirements

• BS, MS, or PhD in Computer Science, Robotics, Electrical Engineering, or a related field.
• 4+ years of experience developing high-performance production software.
• Strong C++ programming skills with experience building, optimizing, and maintaining large-scale software systems.
• Experience optimizing real-time systems, including latency, throughput, memory usage, and multithreaded performance.
• Experience deploying machine learning inference into production environments.
• Familiarity with TensorRT, CUDA, ONNX Runtime, or similar inference acceleration frameworks.
• Strong software engineering fundamentals, including software architecture, testing, and code quality.
• Excellent debugging and analytical problem-solving skills, with the ability to systematically investigate complex software and system-level issues.
• Experience tracing failures across multi-stage pipelines, identifying root causes, and implementing reliable long-term solutions.
• Experience designing validation strategies, automated testing, runtime monitoring, and observability for production systems.
• Strong ownership mindset with the ability to drive problems from investigation through implementation, validation, and deployment.
• Excellent communication skills and experience collaborating across machine learning and systems engineering teams.

Preferred Skills

• Experience with autonomous driving planning, prediction, or robotics software.
• Experience profiling CPU and GPU performance using modern profiling tools.
• Experience deploying software on embedded GPU platforms.
• Familiarity with PyTorch and machine learning workflows.
• Experience with distributed systems, CI/CD, and cloud infrastructure.

Bonus Qualifications

Candidates who stand out typically have one or more of the following:

• Deep expertise in modern C++ (C++17/20) and high-performance software engineering.
• Experience optimizing production inference pipelines for low latency and high throughput.
• Experience developing internal debugging, visualization, profiling, or observability tools.
• Proven ability to diagnose and resolve complex production issues involving machine learning, runtime software, and system integration.
• Strong systems thinking with the ability to balance correctness, performance, maintainability, scalability, and safety.
• Demonstrated technical leadership and experience mentoring engineers on large production software projects.