Designing and developing full-stack SaaS products for clients across logistics, real estate, and e-commerce.

Worked in supply chain operations — this experience directly shaped my ability to build logistics and operations software that solves real workflow problems, not just theoretical ones.

Coordinating campus operations, partner sponsorships, and local youth leadership development events. Structured engagement workflows and public relations campaigns.

Authoring a technical study review documenting antenna signal handoff latency and cell power attenuation parameters in crowded urban network cells.

Organized slide deck demonstrating how logic equations compile into hardware look-up tables on Spartan 6 and Basys 3 boards.

Presented live ESP32 capacitive readings and automated solenoid valve triggers during Dwarkadas J. Sanghvi College's annual projects fair.

Clinical MRI computer vision pipeline. Trains deep binary CNN architectures on brain radiography scans to identify tumor growths and automate triage routing.

Dermatological image diagnostic system. Uses convolutional feature extraction backbones to screen lesion captures and recognize malignant growths.

Biochemical predictive modeling pipeline. Analyzes molecular compound structures to calculate their potential efficiency and recovery rates for treatment.

Smart grid time-series forecasting model. Stacks LSTM recurrent neural layers to model long-term sequential dependencies and forecast regional electricity usage.

Smart city traffic scanning pipeline. Links YOLO boundary object detection with optical character recognition (OCR) engines to capture plate characters from video streams.

High-speed neural constraint satisfaction solver trained on 1 million Sudoku puzzles. Uses 2D spatial convolutions to resolve grid logic.

Real-time driver fatigue CV warning module. Computes real-time Eye Aspect Ratio (EAR) from live webcam streams to trigger alarms.

Self-driving car computer vision tracker. Associates YOLO frame locations with trajectory tracking vectors to map pedestrian movement.

Wearable sensors signal classifier. Processes tri-axial accelerometer and gyroscope telemetry streams to identify human movement sequences.

Symmetric encoder-decoder neural network designed to compress, denoise, and reconstruct image matrices. Fully dockerized.