United States
AI/ML Engineer with 3 years of industry experience, owning 6+ end-to-end production ML systems from problem statement to deployment and monitoring. Founder & President of AI & ML Club at SJSU, leading 64+ officers to deliver large-scale technical events for 2,300+ attendees and hosting senior speakers from Apple, Google, and Microsoft.
Agent Task Planning for Resource-Constrained Client Devices On device AI
- Developed and evaluated edge AI models (400kB–12MB) for computer vision and sound event classification, achieving 93%+ real-world performance (F1) - Owned end-to-end MLOps pipeline (Python, C++) from data collection (80GB+) through model development, deployment, and real-world A/B testing - Led 4 foundational Cortex-M55 embedded AI projects, delivering a demo portfolio and hardware/software optimizations - Advised executive leadership on the strategic integration of ARM Cortex-M55 AI capabilities - Resolved bare-metal hardware and OS-level issues across ARM Cortex-A35 (Linux), M55, and M4 platforms
- Increased sound classification edge AI model performance from 49% to 93% (F1) by curating 29k+ seconds of real-world audio, engineering a data preprocessing/augmentation pipeline, and tuning MFCC features - Designed C++ microcontroller test harness and Python A/B tester to run automated CI/CD edge inference tests, collecting TP/TN/FP/FN metrics under varied real-world conditions - Delivered deployable models on constrained hardware (1.5MB RAM) by quantizing weights (−60% size), and documenting the full stack for team adoption
- Founded and scaled the Club into one of the biggest student organizations at SJSU - Organized events with engineers/leaders from Apple, Google, Microsoft, and Yahoo - Scaled annual attendance from 450 to 2,300 by expanding programming and partnerships - Expanded leadership team from 5 to 64+ officers, establishing structure and role ownership
- Lead a team of 6 in designing and deploying a real-time data collection system on the CAN protocol - Integrated 19 sensors (speed, acceleration, RPM, temperature, suspension, steering wheel, GPS) to optimize vehicle performance through time-series analysis