Developed a graph neural network (GNN) pipeline to detect seizure onset zones from intracranial EEG data of epilepsy patients.

• Reverse-engineered CAN bus signals and optimized C++ firmware to enhance product coverage and data reliability for
fleet telematics device.
• Using SQL and BigQuery pipelines on Google Cloud Compute for automated analysis of vehicle data to diagnose support
gaps from deployed devices, designing Superset dashboards for EV data metrics.
• Developing vehicle signal analysis tools in C# and Python for hardware simulations, time-domain analysis and remote
vehicle access

• Designed high-quality firmware for WLAN ECUs in fully-networked vehicle, integrating secure access point and station
connectivity for seamless OTA updates and driver experience.
• Developed on Android Open Source Project (AOSP) platform using C/C++ (Android NDK) and Java (Android SDK),
adhering to industry-standard automotive software practices.
• Produced comprehensive unit tests using GTest/gMock, JMock, Mockito, and Roboelectric, enhanced error path
handling and increased code coverage to an organizational goal of 90% for critical wireless services.
• Monitored in-vehicle diagnostic telemetry using adb and bug reports to swiftly troubleshoot and rectify Wi-Fi service
crashes.
• Managed and triaged JIRA tickets derived from driver reset log analysis, working with cross-functional teams to
consistently deliver software versions ready for release.
• Leveraged CI/CD best practices using Jenkins for AOSP build automation and GitHub for source control, ensuring
streamlined and efficient software delivery in the automotive domain.

• Developed C++ firmware in Linux environment to deliver real-time welcome/farewell features for in-vehicle infotainment
interface built on Android Automotive OS and QNX.
• Ensured rigorous enterprise software standards were met, utilizing AIDL for seamless feature connectivity and integrating
standardized telemetry, enhancing the networked vehicle experience.
• Transformed high-level design requirements into deployable middleware, leveraging local builds for algorithm verification
with makefile and GNU Make.
• Conducted integration and system testing, flashing images directly onto vehicle hardware, utilizing PuTTY for
diagnostics, and simulating real-world scenarios via CAN bus packet transmission.

• Co-founded a hardware startup automating shingle installation on residential roofs, addressing major industry pain
points with a scalable solution.
• Developed Python algorithms for OpenCV-based live video processing, shingle edge detection, and battery management
(state-of-charge and cell balancing).
• Integrated FreeRTOS and I2C on STM32 MCU for real-time motor control, optimizing shingle placement via camera
feedback and half-duplex communication.
• Conducted hardware-in-the-loop simulations for power, camera, and motor control subsystems before full system
integration.