– Thesis: formal verification of autonomous vehicles and systems

– Additional research: LLM code generation in low resource settings, domain-specific conversational GPT applications

– Implementing scenario video query pipeline to efficiently validate simulation failures against on‐road autonomous driving data

– Developed simulation generation pipeline to convert accident reports involving autonomous vehicles into realistic simulations

– Created NLP framework to infer collision conditions and more robustly evaluate faults in planning and perception algorithms

– Implemented obstacle classification pipeline to forecast severity of road obstacle/debris collisions

– Integrated obstacle detection perception software with real‐time onboard logs to access live lidar data without human assistance

– Integration automates time‐intensive lidar point labeling (saving 8000+ hrs of in‐house labeling)

– Led computer vision R&D project (patent pending) for Lucid Gravity driver monitoring systems

– Created deep learning evaluation platform to detect driver drowsiness or distraction

– IP for infrared system design and implementation under evaluation for next vehicle