Vikas Hanamant Talawar
Bengaluru, Karnataka, India
Full Stack Developer and AI/ML Engineer skilled in Next.js, FastAPI, and MongoDB, building production-ready agentic AI systems and scalable SaaS applications. Experienced in integrating LLM-driven features (Groq Llama 3.1, RAG pipelines, SBERT) to enhance system intelligence. Solved 200+ DSA problems in C++; deployed production apps on Vercel and Render with CI/CD. B.Tech CSE (AI & Data Science), IIITM. GATE 2025 & 2026 qualified. Available for internship or full-time roles immediately.
| Languages | C++, JavaScript, TypeScript, Python, SQL (basics) |
| Frontend | React.js, Next.js, Tailwind CSS, shadcn/ui, Streamlit, HTML, CSS |
| Backend | Node.js, Express.js, FastAPI, REST APIs, Convex |
| Databases | MongoDB, Mongoose ODM, ChromaDB, PostgreSQL (Supabase), Firebase, Firestore |
| AI / ML | Groq AI (Llama 3.1), LangChain, SBERT, RAG pipelines, TF-IDF, scikit-learn, HuggingFace |
| Auth & Tools | Clerk (JWT), Firebase Auth, Zod, PyMuPDF, GitHub REST API, Git, Vercel, Render |
| DevOps | CI/CD (GitHub Actions), MongoDB Atlas, CORS, Chrome Extensions API |
| Core Concepts | Agentic AI, RAG Architecture, DSA, OOP, System Design, Service-Controller Arch |
- Built a scalable real-time chat platform using Next.js, TypeScript, and Convex reactive queries — live messaging, typing indicators, unread tracking, online/offline presence with sub-100ms latency.
- Implemented Clerk authentication securing 100% of API endpoints; architected Firestore schemas for conversations, messages, and read receipts handling concurrent multi-user sessions.
- Developed a Chrome extension for website activity tracking with Chart.js analytics dashboards and Express.js + MongoDB Atlas backend, reducing data retrieval latency by 35%.
- Built a multi-role Student-Teacher Appointment System with Firebase Auth + Firestore supporting Admin, Teacher, Student dashboards across 10+ workflow states with real-time approval and RBAC.
- Developed MovieVerse and Quiz applications using React.js and Tailwind CSS, improving UI interaction responsiveness by 30%.
Python · FastAPI · LangChain · ChromaDB · SBERT · Groq AI · Next.js · TypeScript
- Built an agentic code review system using RAG pipeline with ChromaDB vector store; scored a real codebase 63/100 and identified a critical security issue in production code.
- Semantic code search across 10k+ line codebases using 384-dimensional SBERT (all-MiniLM-L6-v2) vectors with natural language queries over source code.
- Intelligent chunking pipeline reducing API calls by 70% while maintaining full codebase coverage; severity-weighted scoring (critical −15, high −8, medium −4 pts).
Python · SBERT · TF-IDF · Streamlit · HuggingFace · scikit-learn · PyMuPDF
- Fine-tuned SBERT on 853 resume-JD pairs achieving 0.886 Pearson correlation; weighted scoring: semantic (60%), TF-IDF (25%), skill gap (15%) across 73,000+ ESCO skill terms.
- Deployed on HuggingFace Spaces with real-time skill gap visualisation and actionable recommendations.
Next.js · TypeScript · Express.js · MongoDB · Groq AI · Clerk · Zod · Tailwind CSS
- Full-stack AI flashcard generator delivering structured study cards in under 3 seconds via Groq Llama 3.1 with strict JSON output validation.
- 8 typed REST endpoints with Clerk JWT middleware, Zod validation, and service-controller architecture; MongoDB schema with atomic $inc session tracking.
Indian Institute of Information Technology, Manipur (IIITM)
B.Tech in Computer Science Engineering (AI & Data Science)
- CPI: 7.51 / 10
Imphal, India
2022 – 2026
| Introduction to MongoDB | MongoDB University | Jan 2024 |
| Data Analytics with Python | NPTEL | Jan 2025 |
- GATE 2025 & 2026 — Qualified both consecutive years in Computer Science & Engineering.
- LeetCode — Solved 200+ problems: arrays, strings, dynamic programming, and graphs.