Computer Engineering student (AI & ML specialization) @ University of Mumbai β building production full-stack apps and applied AI systems.
- π Final-year B.E. Computer Engineering (AI & ML) student, CGPA 7.8/10 (through 6th semester)
- π 1st Place Winner β Nakshastra APSIT National Level Hackathon (FinTech domain) β built a P2P lending dApp in under 48 hours
- πΌ Actively looking for Full-Stack Development and AI/ML Engineering internships (Mumbai / Navi Mumbai / Pune / Bangalore)
- π§ Comfortable across the stack: React/Node production apps β PyTorch/Transformer fine-tuning β Docker deployment
- π« Reach me at sumit.thakur.dev@gmail.com
SyntaxShift β LLM-based C-to-Python code translator
Fine-tuned Salesforce's CodeT5p-220M on a custom seq2seq pipeline; deployed as an interactive Gradio app for real-time translation.
PyTorch Hugging Face Transformers LLM Fine-tuning Gradio
Fake Aadhaar Detector β Document forgery detection system
Engineered 19 forensic computer-vision features (Error Level Analysis, JPEG artifact analysis, noise estimation) feeding a Random Forest classifier, served via Flask + React.
Python OpenCV Scikit-learn Flask React.js
Coding Nexus β Competitive programming & classroom platform
Production platform with a self-hosted, Dockerized Judge0 judging engine β supports 140+ concurrent users in live college-wide coding competitions, with a teacher dashboard for challenge authoring and real-time leaderboards.
React Node.js Prisma PostgreSQL Docker Judge0
VidVerse β Scalable HLS video streaming platform
Cut initial load times by 35% with an FFmpeg-based adaptive-bitrate streaming pipeline; improved query latency by 40% across 20+ REST APIs using MongoDB aggregation pipelines and Redis-backed background jobs.
Node.js Express MongoDB Redis FFmpeg Next.js
π More projects and experiments: browse all 36 repositories β
Nakshastra APSIT National Level Hackathon β 1st Rank, FinTech domain Architected a decentralized P2P lending dApp (Solidity + React.js) with a decentralized guarantor mechanism, and secured wallet-to-wallet transactions with automated escrow via Ethers.js + MetaMask.
- Machine Learning β Coursera (verify)
Open to Full-Stack & AI/ML internship opportunities β let's connect!


