Full-stack developer & AI/ML engineer building practical computer-vision systems and intelligent web applications.
- 🎓 Education: B.Tech in Computer Science & Engineering at National Institute of Technology, Rourkela (NIT Rourkela, Aug 2025 – Present)
- 💼 Experience: Former Artificial Intelligence Intern at Personifwy / 1Stop AI, leading end-to-end data preprocessing, model architecture design, training, and evaluation for computer vision & NLP systems.
- 🔬 Core Domains: Computer Vision, Multi-Head CNN Object Detection, Transfer Learning, REST API Engineering (FastAPI), React Interfaces & Cryptographic Ledgers.
- 🏆 Hackathons & Recognitions:
- Participant: CodeUtsava (NIT Raipur), Adobe University Hackathon, and HackOdisha 5.0
- Winner: King Hamad American Mission Hospital Medathlon (Dec 2024)
- Winner: Sastra Pratibha, Science International Forum (Apr 2023)
- Certified: Responsive Web Design — freeCodeCamp
- 🌐 Location: Rourkela, Odisha, India 🇮🇳
| Project | Domain | Architecture | Links / Highlights |
|---|---|---|---|
| ChemScope | Molecular Intelligence | React, FastAPI, PubChem API, 3Dmol.js | 🌐 Live App • 💻 Code Interactive 2D/3D chemical graphs & property analysis |
| T-Minus | Aerospace Telemetry | React, FastAPI, Python, Space Devs LL2 API | 🌐 Live App • 💻 Code Rocket countdown telemetry & mission control HUD |
| DigitLocalize-AI | Deep Learning / Object Detection | Python, TensorFlow, Multi-Output CNN, Keras | 💻 Code Joint digit classification & bounding box regression (98.7% acc) |
| BilletVision | Industrial Computer Vision | Python, OpenCV, ArUco, OCR, SQLite | 💻 Code Automated non-contact billet inspection & traceability |
| DeepLandmark | CNN Landmark Recognition | Python, Keras, VGG19, Google Landmarks v2 | 💻 Code Transfer learning classifier across 9,761 landmark classes |
| BlockCert | Cryptography / Blockchain | Python, Ed25519 Signatures, SHA-256 Ledger | 💻 Code Tamper-evident academic credential verification platform |
| HeadlineIQ | NLP Transfer Learning | Universal Sentence Encoder, TensorFlow, AG News | 💻 Code Real-time news topic classification (88.75% test accuracy) |
| PetVision | Breed Classification | ResNet50, ImageNet, Transfer Learning | 💻 Code 23 cat & dog breeds recognized with 86.85% test accuracy |
"Curiosity is a form of creation." • Engineered by Raksha Mendon