My Projects
Click a category to explore.
Hepatoma Staging Pipeline
Constructed an end‑to‑end deep learning pipeline to stage hepatocellular carcinoma using histopathological images. Included segmentation, Stain Colour Normalization, biomarker extraction, and achieved 20% F1 improvement.
Tech stack: Python, Tensorflow, PyTorch, CellProfiler
GAN Fingerprint Detection (ImageCLEF 2023)
Built similarity‑based and autoencoder methods to detect whether real medical CT scans were used to generate synthetic GAN images. Ranked 9th Globally.
Tech stack: Python, PyTorch
Medical Captioning (ImageCLEF 2023)
Developed a gradient-boosted ensemble + deep learning system for concept detection and caption prediction on radiology images. Ranked 7th globally.
Tech stack: Python, TensorFlow, UMLS concept mapping
FungiCLEF (Fungal Species Recognition)
Designed a multi-model ensemble to classify fungal species from microscopic images. Top-tier performance in CLEF‑2022 FungiCLEF.
Tech stack: Python, Tensorflow
SnakeCLEF (Species Classification)
Built an ensemble of convolutional networks to classify snake species from images. Ranked top 10 globally at CLEF‑2022 SnakeCLEF task.
Tech stack: Python, PyTorch, Tree‑based ensemble
WolfJobs
Built a job and resume matching platform for NC State students, incorporating AI-based skill extraction and job ranking. Designed a responsive UI with React and Tailwind, and a backend with Node.js, MongoDB, and Flask. Integrated JWT-based authentication and scalable search.
Tech stack: Python, CellProfiler, U-Net, Ensemble learning
VOGUEX
Developed an ML-powered fashion recommendation engine using DenseNet for feature extraction. Implemented features like login/signup, product reviews, and an image-based search interface with Flask backend and SQLite storage.
Tech stack: Python, Autoencoders, SSIM, Patch-based methods
HAART
Designed and deployed a full-stack web application for anomaly detection in heart rate data using time series classification. Built a Grafana-based dashboard for real-time monitoring and alerts, fully containerized with Docker.
Tech stack: Python, TensorFlow, XGBoost, UMLS concept mapping
Google Data Analytics Capstone
Analyzed customer data from a bike-share company to derive actionable marketing insights. Cleaned and transformed 12M+ rows using R, visualized usage trends in Tableau, and delivered strategic recommendations.
Tech stack: R, Python, Tableau
Quantitative Funding Curve (QFC)
A Monte Carlo backtesting engine for shorting strategies on Hyperliquid using historical candlestick data. Optimizes entry/exit conditions across millions of parameter combinations to maximize portfolio returns.
Tech stack: Python, Hyperliquid API
Pet Projects (High School)
A collection of beginner-level fun projects including a Discord bot, a Railway and a Hotel Management System. (My First few projects in High School !!)
Tech stack: Python, C++, Java, JavaScript, APIs