John Paul Hilson
- AI/ML engineer specializing in computer vision, generative AI, and retrieval-augmented generation — turns research into working applications, not just proofs of concept.
- Published research at ICDAM-2025 (Springer) and accepted at ICICC-2026; hands-on industry experience through an internship at Tech Mahindra.
- B.E. in Computer Science (AI & ML), Chandigarh University, graduating 2026 — open to AI/ML engineering roles and research collaborations.
Bill of Materials
I build AI systems the way I'd design an experiment — hypothesis-driven, measured, and reproducible enough that someone else could pick up where I left off.
During my internship at Tech Mahindra, I worked on computer vision for an MBSE-based automated guided vehicle proof-of-concept and contributed to retail and cold-chain frameworks — where "it works in the demo" and "it works on the warehouse floor" turned out to be very different bars to clear.
Outside of coursework, I've had two papers accepted — one on AI-driven fashion curation (Springer, ICDAM-2025) and one on emotion and eye-response fusion for mental wellness (ICICC-2026).
| No. | Description | Category | Level |
|---|---|---|---|
| 1 | Python | Languages | |
| 2 | PyTorch / TensorFlow | Deep Learning | |
| 3 | Computer Vision (OpenCV, MediaPipe, YOLO) | AI/ML | |
| 4 | RAG / FAISS / Sentence Transformers | GenAI & Vector Search | |
| 5 | LLMs / Prompt Engineering | GenAI | |
| 6 | SQL / Streamlit | Web & Data |
Selected Projects
AI-Powered Exam Question Generator
Built an AI application that generates syllabus-based exam questions from uploaded PDFs using Retrieval-Augmented Generation. Implemented semantic search for context retrieval and AI-assisted document chat, with customizable question types (MCQ, short/long answer, true/false, fill-in-the-blank) and PDF/DOCX export.
View Assembly →Emotion & Eye-Response Fusion for Mental Wellness
Deployed a lightweight, CPU-friendly interface fusing facial emotion recognition and eye-tracking to classify Relaxed, Focused, Stressed, and Fatigued states, with real-time adaptive feedback generated via LLMs. Accepted at ICICC-2026.
View Assembly →AI-Driven Personalized Fashion Curation
Developed a virtual fashion try-on concept using deep learning, generating clothing recommendations based on skin tone, gender, and body shape ahead of a virtual try-on step. Published in ICDAM-2025, Volume 6, Springer.
View Assembly →Career Revisions
| Rev | Date | Description of Change |
|---|---|---|
| A | Jul 2025 — Dec 2025 | Support Trainee,Tech Mahindra Built proof-of-concept frameworks for retail and cold-chain supply problems; conducted applied research for kitchen appliance systems; contributed computer vision components to an MBSE-based automated guided vehicle proof-of-concept. |
Request a Consultation
Prefer to skip the form? I read every message that lands directly, too.