Ongoing - September 2025 - Present
Engineered smart soldering station with integrated fume extraction, automated PCB positioning, and adaptive lighting, reducing setup time by 60% and improving workspace ergonomics. Implemented real-time hand gesture recognition using OpenCV and MediaPipe, enabling touchless control of 5 station functions with 94% gesture classification accuracy.
January 2025 – April 2025
Designed 5-stage pipelined MIPS processor in Verilog (IF/ID/EX/MEM/WB) supporting 30+ R-type, I-type, and J-type instructions with full hazard detection and data forwarding logic. Implemented branch predictor and forwarding unit, reducing pipeline stalls by 65% and achieving 3.8x speedup over single-cycle baseline on sorting and matrix multiplication benchmarks.
March 2024
Architected a full-stack web application using a React frontend and Flask backend that leverages OpenAI GPT-4 API to provide students with personalized, AI-driven course registration roadmaps. Deployed application on AWS (EC2, RDS, S3) with 99.9% uptime, winning 1st place at HackPSU Spring 2024 among 50+ competing teams for technical execution and user impact.
Ongoing Research
Developed Python-based data acquisition interface for electrochemical glucose biosensors, achieving 1000+ samples/hour throughput with 99.3% accuracy using custom signal processing algorithms. Implemented LSTM-based predictive model using TensorFlow to forecast sensor degradation with 89% accuracy, enabling proactive maintenance and extending sensor lifetime by average of 2 weeks.
Final Project Demo
Complete demonstration of the smart soldering station with gesture recognition, automated positioning, and integrated fume extraction system