Professional Experience

Feb 2026 – Present

Leading end-to-end autonomous vehicle and robotics solutions, architecting systems that balance latency, safety, and compute constraints across sensors, perception, and decision-making
Exploring VLA models including NVIDIA Alpamayo for advanced robotics perception and autonomous vehicle vision tasks, with simulations in CARLA, IsaacSim and Omniverse
Leading the Vision Pilot project within the Autoware Foundation, driving the transition from modular architectures to end-to-end systems, with responsibility for model development, dataset curation, pipeline optimization, and production releases.

2024 – 2026

Contributing to Autoware Foundation on real-time obstacle perception, working on vision-based segmentation, depth estimation, and path prediction for autonomous driving. Focused on training, data auditing, model quantization via ONNX Runtime for real-time inference.
Built LLVM-MCToll pipelines for static ARM↔️x86 binary translation; validated APKs in Android Studio
Developed Linux kernel modules for ACPI device enumeration and debugged low-level I2C/UART drivers

2024

Developed LLM-guided motion planning that reduced computational nodes by 27.1%
Built multi-agent architectures for high-dynamic vehicle maneuvers using physics-informed control
Improved maneuver success rates to 90% through simulation validation of LLM-planner outputs

2023 – 2024

Deployed ROS-based robotics infrastructure using Docker/Podman containers and Grafana dashboards
Managed secure access via Keycloak and improved CI/CD for real-time robotic system updates
Migrated robotic backend services and monitored system health using custom metrics pipelines