Controls-first engineer applying robotics to real vehicles.
I design and ship control algorithms and autonomy plumbing — EKF/LQR/MPC, sensor fusion across IMU/GNSS/wheel-speeds, and ROS2 deployments on NXP S32G & Raspberry Pi. Strong MATLAB/Simulink + Python, with CAN, I²C and Docker in the loop.
Highlighted Projects
Vx/Vy EKF with Slip‑Aware Fusion
Nine‑state EKF fusing wheel speeds, IMU yaw‑rate/accel, and GPS velocities. Dynamic noise scaling under slip, low‑speed yaw freeze, delayed measurement handling via buffer replay, and bias estimation.
Lateral & Longitudinal Control
Lateral control to manage rollover risk using a rollover index and lateral‑accel limits; steering impulse shaping. Longitudinal control for accel/decel, regeneration blending, power/ range estimation, and calibrated steering map tracking along waypoints.
PointPerfect NTRIP Client (ROS2)
TLS client that discovers mountpoints, sends live GGA, and streams SPARTN/RTCM3 corrections to u‑blox receivers. Auto‑reconnect and metrics heartbeat; integrated with ROS2/Docker for cm‑level positioning.
Emergency Vehicle Detection (Audio ML)
Binary siren detector via YAMNet embeddings with a custom TFLite head. Real‑time ROS2 node publishes detection status and amplitude, with advert/alert duck/restore logic.
Mass & Grade Estimation
Vehicle mass and road grade estimation using Recursive Least Squares with torque, road‑load, and acceleration signals; tuned for <10% error on test routes.
CAN Telemetry + Web Dashboard
SocketCAN receivers to ROS2 topics with a lightweight dashboard for GPS fix, wheel speeds, IMU, and state. Dockerized with CI/CD for edge deployment.
Experience
Magna International — Sr. Software Engineer (Apr 2024 – Present)
- Built EKF‑based estimation for Vx/Vy, heading, and pose from IMU, GPS, and wheel speeds; added IMU line‑cal via I²C and roll/pitch (Kalman & Madgwick).
- Designed lateral control with rollover index / lateral‑g limits and steering impulse; longitudinal accel/decel & regen blending; range estimation.
- Integrated moving‑base GNSS corrections via NTRIP into ROS2/Docker; built audio pipeline (alerts/ads) and ML siren detection.
- Mass/grade RLS estimation with ~10% error margin on routes.
Dana Incorporated — Technical Lead (Jan 2023 – Apr 2024)
Led AUTOSAR ASPICE‑compliant ASW for BEV transmissions; managed 6 engineers, technical reviews, OEM workshops.
FEV North America — Sr. Controls / Technical Lead (Oct 2018 – Jan 2023)
Torque‑path controls: creep, dual‑path shift energy, coast‑based split, launch control, predictive control enhancements. LQR‑based engine start/idle control profiling; regen‑brake interfaces for smooth shift transitions.
Skills
Controls: Kalman Filters, Particle Filter, LQR/LQI, MPC, Optimization, Feedback Control
Software: MATLAB/Simulink, Python, ROS2, Docker, C++, JSON/YAML
Hardware: NXP S32G, Raspberry Pi, DSpace, ETAS/INCA, MotoHawk, CAN, I²C/UART
Tools: Git, Azure, CI/CD, AUTOSAR, Agile, Polarion, Doors