Perception Integration Software Engineer

### Who you are
- Strong proficiency in modern C++ (C++14/17 or newer)
- Production experience with state estimation and multi-sensor fusion, including EKF/UKF, motion models, data association, and uncertainty handling
- Experience designing metrics and evaluation frameworks for real-time, ML-driven robotics systems
- Strong debugging skills across distributed and asynchronous pipelines, using logs and metrics to drive root-cause analysis
- Excellent verbal and written communication skills

### What the job involves
- The Glydways perception team is responsible for designing and implementing a perception system that includes data fusion from state-of-the-art sensors (e.g., LIDAR, RADAR, High definition cameras, Ultra-wide-band radios, etc.) and robust detection, classification, and tracking of any and all obstacles that could present a hazard to the vehicle (Glydcars) in the system
- The perception system for Glydways will reason about information from each Glydcar and from regularly spaced sensor pods monitoring the road network
- The perception team plays a vital role within the Autonomy Software engineering team. The team’s deliverables include:
- Documenting a detailed design for enacting a safe system that can be certified by public transportation authorities
- Implementing said design on mature prototype vehicles in demonstrations to customers, potential customers, and investors
- Expanding the design to include fail-operational capabilities that extend the overall system to safely give our customers a more comfortable experience
- Managing data collection, autonomy testing, and milestone demonstration events that showcase the system’s maturing capabilities, leading to a production system
- Team members may work remotely and must be self-motivated
- Design, implement, and maintain high-performance C++ modules at the boundary of detection like tracking, and prediction
- Integrate outputs from multiple sensors(camera, lidar, radar, map) into EKF-based tracking and fusion systems
- Own metrics, evaluation pipelines, and tooling to measure detection and tracking quality and catch regressions at scale
- Define and maintain clear interfaces between perception, tracking, and planning modules with a focus on robustness and debuggability
- Develop simulation, log-replay, visualization, and offline analysis workflows to improve test coverage before on-vehicle deployment