Welcome to my GitHub! I'm currently pursuing my B.Tech in Electronics and Communication Engineering at NSUT (Class of 2027), but my true passion lies in code. I bridge the gap between heavy-duty algorithms and sleek full-stack applications.

Instead of just building standard web apps, I love diving into the deep end of GPU-accelerated inference pipelines, hybrid recommendation engines, and robust authorization frameworks.

• 🧠 AI-Powered Systems: Deploying low-latency deep learning models and managing unstructured datasets.
• ⚡ Competitive Programming: I'm constantly sharpening my problem-solving skills. I've tackled over 650+ problems on LeetCode (Rated 1710) and actively compete on Codeforces (Specialist - 1453) and CodeChef (3-Star - 1627).
• 🏗️ Type-Safe Full-Stack: Building out complex, scalable web architectures using tools like Next.js, FastAPI, and PostgreSQL.

• Languages: C++, Python, TypeScript, JavaScript, SQL
• Frontend & Backend: Next.js, React.js, FastAPI, Node.js, Express.js, tRPC
• Databases & Infrastructure: PostgreSQL, MongoDB, Drizzle ORM, Cloudflare R2, Modal, Docker
• Machine Learning: PyTorch, Scikit-Learn, SciPy, Librosa, RapidFuzz, DVC

A low-latency Text-to-Speech platform. I engineered a GPU-accelerated deep learning pipeline using Modal for the model inference and set up Cloudflare R2 to efficiently handle all the unstructured audio data.

A recommendation engine that actually understands the music. It merges acoustic analysis with collaborative filtering. I used DVC to track a pipeline that processes over 600MB of data into SciPy sparse matrices, exposed via a fast FastAPI backend with fuzzy searching.

A highly scalable authorization system. I implemented both Role-Based (RBAC) and Relationship-Based (ReBAC) access controls using CASL, backed by a relational database schema designed with PostgreSQL and Drizzle ORM.