> Markdown version of [/jobs/ext/4607561-autonomy-zip-perception-swe](https://www.wearedevelopers.com/jobs/ext/4607561-autonomy-zip-perception-swe). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Autonomy Zip Perception SWE

- **Company:** Zipline
- **Location:** Ann Arbor, MI, United States
- **Experience:** Expert
- **Salary:** $200,000.0 - $240,000.0
- **Contract:** Permanent contract
- **Skills:** Computer Vision, Software Debugging, Machine Learning, Object Detection, Visual Systems, Deep Learning, Reliability of Systems, Low Latency
- **Published:** November 6, 2026
- **Apply:** https://www.adzuna.com/details/5852209814

## About the Role

* At least 5+ years (Senior) or 8+ years (Staff) of professional experience crafting and deploying sophisticated perception systems for autonomous robots, aircraft, or vehicles.
 * Expertise in state-of-the-art deep learning models, specifically focused on object detection and tracking, honed through real-world deployments.
 * Deep knowledge of geometric computer vision techniques, including stereo vision, depth-from-motion, structure-from-motion, and visual odometry, for precise depth and 3D position estimation.
 * Innovative mindset and practical problem-solving skills, demonstrated by a proven track record of overcoming performance bottlenecks and resolving operational edge cases.
 * Hands-on experience supporting commercially operated autonomous systems, including diagnosing issues and enhancing system reliability in live environments.
 * Exceptional communication and teamwork abilities, capable of thriving in a dynamic, cross-disciplinary engineering environment committed to safety-critical operations.

## Description

* Engineer Reliable Vision Systems: Develop scalable model training pipelines and build rigorous evaluation frameworks that capture real-world performance across edge cases and long-tail scenarios.
 * Integrate Across Systems: Collaborate with systems, flight test, and validation teams to design end-to-end test plans that push perception to the edge of capability - and beyond Zipline's strict safety and reliability standards.
 * Own Operational Performance: Create powerful debugging, visualization, and analysis tools that drive insight from field data, enabling rapid triage and root-cause analysis of perception issues in live deployments.
 * Learn Fast, Iterate Faster: Leverage simulation and real-world feedback to continuously refine both ML models and classical tracking algorithms - improving accuracy, latency, and robustness in challenging operational conditions.
 * Set the Standard for Airspace Autonomy: Contribute to the long-term vision and architecture of Zipline's detect-and-avoid system, helping shape a perception stack that leads the industry in safety, reliability, and autonomy.