Nipun Chugh

I’m a BE in Computer Science student who enjoys turning complex ideas into working products—especially where NLP, ML, and strong engineering meet. I build with Python and modern web stacks, and I like shipping features end-to-end, not just prototypes.

In my FlashCard Engine, I built a full-stack FastAPI engine that auto-generates flashcards from PDFs using Groq’s Free API (Llama 3:70b) and implemented the SM-2 spaced repetition algorithm with a 4-button rating system and minute-based lapse recovery. I also created async PDF processing with ThreadPoolExecutor and frontend polling, accelerating flashcard generation 3x, and designed a multi-provider LLM architecture with JSON parsing, fallback logic, and bcrypt-secured sessions.

I also developed an AI Resume Analyzer + Job Matcher that extracts skills with spaCy NLP and scores resumes across five categories, recommending roles across 15 job types using TF-IDF and Cosine Similarity. For ML, I created EmPath-AI, a multi-task DistilBERT pipeline that classifies 28 emotions and detects crisis-related signals with a DistilBERT-based approach, integrating safety-filtered generative logic for responsible mental health support; and I built a Brain Tumor Detection CNN pipeline with PyTorch, OpenCV, and Python, achieving 92% validation accuracy and 88% sensitivity.