• Developed real-time robust frameworks for autonomy compute communication (sensing, compute, and controller inter-ECU communication).
• Developed a real-time EDR that records critical AD events with images and contextual logs for an L2-EV enabled with tri-focal cameras and MobileEye (EQ4) vision chip.
• Was invited as a speaker to talk on EDR - state of the art at Wayne State University's Graduate Research class.
• Developed an automated validation framework for camera calibration by emulating simultaneous image and message replays to the autonomous domain (AD) controller.
• Lead the development of the 2-wire BroadR-Reach Automotive Ethernet with NxP TJA1100 PHY and Marvell 88Q5050 switch. First in the company to deliver a full stack time-triggered real-time communication for Electrified Powertrain (ePT) Network.
• Architected, designed, and implemented a generic Inverter State Machine (ISM) to support inverter application behavior when interacting under normal, standby, fault, and limp states.
• Lead engineer responsible for requirements, design, development, and successful delivery of a generic application platform for DCDC 800/400 within a short turnaround time of 4 weeks.
• Responsible for architecting other platform features such as multicore IPC, EEPROM, PWMs using GTM (Generic Timer Module), Emergency Stop (ESTOP), VADC support for Current Sensors, Signal Qualifier Module (SQM), LIN stack with python auto code for message serialization and deserialization.
• Evangelist for tools development and automation within the ePT team. With an intern developed a $25 sniffer tool for CAN and two-wire Ethernet using COTS like Raspberry Pi, PiCAN2 shield, Android tablet, and opensource Python libraries (cantools, socketCAN, pyshark). Established automotive tool vendors charge over $25000 per seat license for such tools.