AI/ML Engineer  ·  GenAI Developer  ·  Software Engineer

I build production-grade AI systems — RAG pipelines, LLM applications, and intelligent full-stack products that solve real problems.

• 🎓 B.Tech CSE @ Quantum University — CGPA 8.3 | Graduating 2026
• 🧠 Specialized in Generative AI · RAG · NLP · Computer Vision · LLM Applications
• 💻 821+ LeetCode problems | Global Rank 46,622 (Top 6%) | Rating 1561
• 🏆 Top 7 Finalist — HackIndia Hackathon  |  2nd Place — Genesis Idea Open House 2023
• 🚀 Actively seeking AI/ML Engineer / GenAI Developer roles (Full-time & Internship)

Programming

ML / DL Frameworks

GenAI / LLM

Data Tools

Dev & Web

Problem: Extracting accurate, cited answers from multiple PDFs at scale with low latency.

FastAPI LangChain FAISS Llama-3-70B (Groq) Cohere Embeddings Next.js

• 📈 Retrieval accuracy: 68% → 84% (validated on 150 real-world queries)
• ⚡ Latency: 3.1s → 1.5s via Groq inference + async FastAPI + embedding caching
• 🛡️ Citation grounding + fallback responses to mitigate hallucinations
• 🔧 Pipeline: PDF ingestion → 800/150-token chunking → embeddings → FAISS → MMR retrieval → LLM synthesis

🔗 Live  |  📂 GitHub  |  🎬 Demo

Problem: Single-model recommenders fail to capture visual + behavioral user preferences together.

PyTorch MobileNetV2 FastAPI Scikit-learn

• 🎯 88% validation accuracy combining collaborative filtering, metadata similarity, and visual feature extraction
• 📷 Applied Computer Vision (MobileNetV2) for image-based item matching
• 🏗️ Modular, testable backend API architecture — scalable to production

Problem: Inconsistent revision and poor long-term retention of DSA concepts during interview prep.

React.js Node.js MongoDB JWT

• 📊 Revision consistency improved by 40% via spaced-repetition + decay scheduling
• 🗂️ 500+ problems logged · 20+ active users · personalized modes (by company, topic, random)
• 🔐 JWT auth with refresh token rotation · daily streak tracking · analytics dashboard

🔗 Live  |  📂 GitHub  |  🎬 Demo

Problem: Discovering similar movies based on plot, genre, and cast — fast and accurately.

Python Scikit-learn TF-IDF NLP Streamlit TMDB API

• ⚡ 60% faster lookups via precomputed similarity matrix + NumPy vectorization
• ✅ Validated across 50+ real-world searches
• 🎨 Interactive Streamlit app with live TMDB poster previews

🔗 Live  |  📂 GitHub

🌐 Web Content Intern — GAO Tek Inc. Aug 2024 – Oct 2024 · Remote

• Optimized 100+ product listings for SEO using metadata and keyword analytics

Open to AI/ML, GenAI, and Software Engineering roles · Full-time & Internship · 2026