Sep 2024 – Present · 1 yr 9 mos · On-site

• Owned USS (Ultrasonic Sensor System) service development end-to-end in C++ using SOA architecture for production automotive ECUs — real-time diagnostics, fault handling, and ASIL-aware behavior across 5+ integrated services.
• Independently designed, mapped, and implemented 250+ USS diagnostic faults on A72 and R5 cores, derived from SYS1 and SYS2 system requirements — covering fault detection, reporting, and handling across the full USS diagnostics stack.
• Served as the primary technical interface between client and the external USS algorithm supplier — managing third-party library integration, aligning on interface contracts, and coordinating issue resolution across teams.
• Debugged and resolved integration issues at the boundary between Client platform and the external algorithm library — including interface mismatches, signal mapping errors, and algorithm initialization
• Drove CPU utilization from ~70% to ~30% (~57% reduction) through execution profiling, memory optimization, and resource handling on a safety-critical embedded target.
• Reduced io-sock CPU usage by ~20% by overhauling QNX socket handling and eliminating redundant socket creation.
• Worked on Autonomous Parking ADAS stack integration across 6 modules — Environment Reconstructor, Mode Manager, Map Horizon Interface, Slot Planner, Behavior Planner, and Motion Control.
• Resolved complex production crashes and concurrency bugs across QNX aarch64 and SafeRTOS R5F using GDB and static analysis — heap corruption, race conditions, IPC failures, and CPU spin issues.
• Diagnosed and fixed timesync, cross-core IPC, and sensor fusion issues across multi-core heterogeneous SoCs (A72/R5F) on TI J784S4.
• Integrated embedded AI inference pipelines (TIOVX/TIDL/OpenVX) on TI Jacinto for ML-enabled ADAS features.
• Maintained OTA deployment pipelines via Wind River Studio; contributed to ECU bring-up, boot-time optimization, and production validation.

Sep 2022 – Sep 2024 · 2 yrs · On-site

• Completed training in Linux, C++, C, Python, and AI fundamentals, supplemented by AI projects.
• Integrated an ADAS stack into CARLA simulation using ROS, conducting functional testing within virtualized environments to simulate real-world scenarios.
• Worked on an ADAS stack for a client, implementing a Service-Oriented Architecture (SOA) leveraging SOME/IP, TCP, and shared memory for seamless service communication.
• Conducted performance evaluation across real-time and embedded operating systems, ensuring optimal execution on industry-standard platforms.
• Assisted in hardware-software integration, supporting testing procedures to ensure reliability on diverse automotive-grade hardware.
• Served as an integrator and tester, ensuring seamless builds and deployments using a CMake-based system while collaborating with cross-functional teams.