I have gained hands-on experience with NestJS, particularly its approach to dependency injection and inversion of control. I’ve also worked on projects such as a Library Management System and an Appointment Management System, which helped me strengthen my development skills and apply concepts in real-world applications.
Taught 6th-grade students foundational programming concepts, including an introduction to Python and the basics of AI and Machine Learning, fostering early interest and understanding in technology and innovation.
Gained hands-on experience with hotel infrastructure by maintaining PCs, network devices, and cabling, ensuring smooth operation and connectivity across the facility
Graduation project: designed and developed a multimodal stress detection system leveraging VR telemetry, pose estimation, and deep learning. Created a VR-based data collection environment, processed behavioral and motion data, and trained time-series models to classify stress in a passive and scalable manner.
Dummy FastAPI server project for practicing MLOps, Kubernetes, Docker, and CI/CD pipelines. Implemented a pet portrait printing service with image processing and model deployment.
Developed an XAI-based diagnostic system achieving 72% average accuracy on the HAM10000 dataset using PyTorch. Implemented CNN architectures including ResNet50, Inception, and custom CNN networks, and integrated explainability methods to improve interpretability for dermatologists.
Implemented and evaluated a retrieval-augmented generation system from scratch, reducing LLM hallucinations.
Engineered a CNN-based detection system achieving 99% accuracy on a custom dataset of 50,000 images. Implemented ensemble learning with Inception-v3 and ResNet-50, reducing false positives by 30% compared to single-model approaches.
Implemented the Adagrad optimization algorithm from scratch as a TensorFlow optimizer for a linear and nonlinear programming course.
Built a machine learning project for classifying user activities based on accelerometer and gyroscope data. Included exploratory data analysis with Pandas and Matplotlib, and used KNN, SVM, Logistic Regression, and TensorFlow neural networks.
Built a cross-platform, local-first React Native app for face recognition-based attendance tracking. Powered by FaceNet with TFLite and ONNX frameworks using a Java implementation, and stores attendance data in a SQLite-based vector database.
Developed a real-time Android application for door and obstacle detection using TensorFlow Lite and YOLOv11. Integrated CPU, GPU, and NNAPI hardware acceleration with audio feedback for accessibility, and added configurable detection parameters for visually impaired users and navigation assistance.
Designed and developed an interactive Tableau dashboard to track personal fitness data, including weight progression, top lifts, reps by weight, cardio distance, and workout frequency. Enabled performance analysis and trend visualization to support goal setting and consistency in training.
Built a data visualization dashboard for analyzing company performance trends over time using Python, Flask, and AmCharts.
Built a full-stack software engineering project using Python and Flask for the server, PostgreSQL, Firebase push notifications, Next.js, Docker, and CI/CD pipelines.
Built a web scraper that converts schedules from Ellucian Self-Service into ICS files for seamless integration with calendar providers. Uses Puppeteer to scrape schedule data from the Self-Service dashboard.
Reverse engineered a smartwatch to retrieve heart-rate data on a PC. Used Bluetooth Low Energy and the Generic Attribute Profile (GATT) to read data, and created a cross-platform application with Electron, HTML, CSS, Bootstrap, and JavaScript.
Open-source cross-platform Electron app for creating dynamic video slideshows. Users can bind text slides to video backgrounds using Node.js, HTML, and CSS.