Senior Motion Planning Software Engineer - Droid

Zipline•South San Francisco, CA

3d•$200,000 - $245,000

About The Position

Zipline operates the world’s largest autonomous logistics system (ground or air) — and in the coming year, we will be hyper-scaling from thousands to tens of thousands of carefully coordinated drone deliveries in several dense, dynamic U.S. metros. Our Autonomy Motion Planning team is looking for a Senior Software Engineer who is passionate about developing autonomous systems for the real world. This role will explore cutting-edge approaches to decision making and trajectory planning that will enable more accurate and timely deliveries, putting you in the position to make critical product decisions that will shape our future architectures.

Requirements

Master's degree in Computer Science or related field and 3+ years of experience building software for safety-critical systems (aerospace/AV/robotics).
Strong in Rust/C++/C for real-time, fault-tolerant code on embedded/Linux.
Hands-on with simulation at scale, SIL/HIL, log replay, and metrics-driven validation.
Evidence of shipping production-grade autonomy through ambiguous, noisy conditions—owning the last mile to reliability.
Systems thinker who collaborates tightly with perception, controls, and flight ops; crisp docs and design reviews.

Responsibilities

Build real-time trajectory generation and decision-making for autonomous flight (search-based, sampling, MPC, convex/non-convex optimization).
Tackle joint optimization across safety, energy, time, and reliability—balancing mission goals with vehicle constraints.
Plan in uncertain environments with complex dynamics: robustness to wind/turbulence, degraded sensors, and partial observability.
Design safety-critical software for avoiding obstacles and interfacing with humans while delivering and picking up packages
Extend the autonomy stack for new aircraft and payloads; define clean interfaces with perception and controls.
Prove it before flight: large-scale sim, SIL/HIL, log replay, and fault-injection.
Mine real fleet data to validate safety metrics, improve models, and burn down long-tail failure modes.