AeroVect - Principal Engineer, Autonomy

South San Francisco, California , US$300k - $350k+ Equity3w ago

In Office Principal NA Oil & Gas Airlines Principal Engineer C++Python Team Leadership Mentoring Close

Requirements

• 15+ years of hands-on experience building production autonomy systems, with strong technical depth across multiple modules (localization, perception, prediction, planning, controls). You think at the level of the autonomy system, not a single module. • Demonstrated track record of shipping autonomy components that have run in production on real vehicles at non-trivial scale — not just research prototypes or simulation results. • Prior experience as the most senior individual contributor in an autonomy organization — setting direction, mentoring staff/senior engineers, and partnering with engineering leadership without managing a team yourself. • Deepest technical depth in perception, prediction, or planning (ideally more than one of the three). • Strong software engineering fundamentals in C++ and Python. You write or review code that other senior engineers want to extend and trust in a safety-relevant system. • Fluency with modern deep learning for autonomy, including the practical realities of training, evaluation, deployment, and lifecycle management of models that have to work in the real world. • Experience working in or with ROS / ROS 2 and the distributed-systems realities of on-vehicle compute (real-time constraints, IPC, fault containment). • A bias for execution. You ship. You close out problems. You convert ambiguity into a plan and the plan into running code on a vehicle. • We Prefer • Experience with safety-critical or functional-safety-relevant systems (ISO 26262, ISO 13849, SOTIF, or aerospace equivalents). • Experience operating in an Operational Design Domain that involves heavy interaction with humans, mixed traffic, or unstructured environments. • Familiarity with simulation-driven verification and the use of simulation as part of a CI/CD pipeline for autonomy.

Responsibilities

• You are the senior-most IC in autonomy, with the deepest technical ownership of either Perception, Prediction, or Planning (or any combination thereof) and influence across the rest of the stack. • Perception, Prediction, • Planning • Perception. Own the design and evolution of the perception stack — detection, classification, tracking, and multi-modal sensor fusion across the available modalities. Drive perception robustness across the long tail of real-world operating conditions, and set the direction for where and how deep learning is best applied across the perception pipeline. • Perception. • Prediction. Own the prediction stack and the design of models for intent inference, behavior forecasting, and handling occlusions and edge cases. Set the direction for how prediction integrates with perception upstream and planning downstream. • Prediction. • Planning. Own the design of the planning and decision-making stack, from structured driving behaviors to the domain-specific maneuvers required for autonomous GSE operations. Set the direction for where learned components earn their place in the planner. • Planning. • Cross-stack influence. Set the technical direction at the interfaces between your primary areas and the rest of the stack, and partner with the other senior engineers in autonomy to keep the system coherent end-to-end. • Cross-stack influence. • Autonomy architecture. Own the functional and SW architecture of the autonomy stack, and partner with neighboring teams towards its implementation. • Autonomy architecture

Benefits

• San Francisco$300K – $350K • Offers Equity • A real ODD with real constraints. Airports are one of the few environments where commercial autonomy is genuinely viable today and where the path to removing the safety driver is concrete rather than speculative. • A real ODD with real constraints. • Scope. This is the senior autonomy IC role at AeroVect. You are the senior-most technical voice across Perception, Prediction, and Planning, with cross-stack influence across the rest of the autonomy stack. • Scope. • A defined path to scale, not a science project. A real commercial deployment with a concrete path to removing the safety driver and scaling the fleet. Your work has a destination. • A defined path to scale, not a science project.