Software Engineer, Localization

Mach Industries • San Francisco, CA

1d

About The Position

Mach Industries is building an AI-forward autonomy stack for contested environments where GPS and other sensing may be unavailable or unreliable. The Software Engineer, Localization will design and implement estimation and sensor-fusion algorithms for robust navigation across the company’s product lines. The work spans perception, state estimation, and embedded systems, taking research-grade algorithms into rugged real-world deployments. The role involves prototyping and productionizing vision navigation and targeting features, integrating sensors, optimizing real-time software, and validating performance through simulation, hardware-in-the-loop testing, and flight testing. The position may involve access to information controlled under U.S. export regulations.

Requirements

• Modern C++ on Linux and Python for tooling and analysis, with strong debugging, profiling, and testing discipline.
• SLAM and state-estimation experience, including error-state EKF/UKF, factor graphs, nonlinear least-squares methods such as Ceres or GTSAM, observability, and covariance tuning.
• Vision experience with VIO/SLAM, camera models, optical flow, and feature tracking.
• Comfort with deep learning for detection, segmentation, or pose estimation using PyTorch, and with on-edge deployment.
• Sensor-integration experience with IMU strapdown and biases, GNSS/RTK, and multi-camera, LiDAR, radar, magnetometer, and barometer systems.
• Proven delivery of research-to-production systems and field testing on real platforms.
• Five years of experience, with a BS, MS, PhD, or equivalent practical experience in Computer Science, Robotics, Electrical or Aerospace Engineering, or a related field.
• Must meet applicable U.S. export-control eligibility requirements or be eligible to obtain required authorizations.

Nice To Haves

• Experience with CUDA, TensorRT, or ONNX Runtime and NVIDIA Jetson pipelines.
• Exposure to ROS 2 and PX4/ArduPilot integration.
• Strong data practices, including data validation in CI, SQL/Parquet, and reproducible datasets.
• Experience in contested or denied RF, low-light/night, or high-vibration environments.
• Rust for systems tooling and Docker for reproducibility.

Responsibilities

• Prototype and productionize vision navigation and targeting features end-to-end, from simulation through hardware-in-the-loop testing to flight, using production C++.
• Turn detections from EO/IR, RF, and radar into well-posed measurement models with latencies and covariances, and make the estimator decision-aware without corrupting state.
• Stabilize GNSS-to-VIO handover using adaptive covariances, gating, hysteresis, and reset-less alignment to eliminate jumps and estimator resets.
• Build and optimize real-time software on Linux and embedded platforms; profile CPU/GPU performance and vectorize hot paths, with optional CUDA/TensorRT work on Jetson hardware.
• Own calibration and time synchronization across IMUs, cameras, radar, LiDAR, and GNSS, and validate in flight.
• Create evaluation pipelines and dashboards for drift, handover stability, relocalization, and track quality.
• Implement fault detection and graceful degradation for harsh conditions such as blur, low light, vibration, and RF denial.
• Integrate global aids such as maps, magnetics, and radar for long-term consistency and loop-closure robustness.

Benefits

• Performance-based incentives tied to individual and company performance.
• Stock options or equity participation.
• Healthcare, dental, and vision plans.
• Retirement savings plans.
• Paid time off.
• Funds for continuing education, training, and career growth.

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level