//
CS undergrad building AI-powered products and full-stack web apps — from RAG systems and LLM pipelines to production-ready REST APIs and SaaS dashboards.
// 01 — About
I'm a Computer Science undergrad at Hindustan College of Science & Technology, Mathura, passionate about turning raw data and large language models into products people actually use.
My work spans both the ML lifecycle and full-stack engineering — from data preprocessing and LLM pipelines to building RESTful APIs, SaaS dashboards, and deploying production apps end-to-end.
Currently exploring the intersection of GenAI and practical software engineering: RAG pipelines, semantic search, LLM-powered tools, and scalable backend architectures with React & Node.js.
// 02 — Skills
// 03 — Experience
// 04 — Projects
Architected a GenAI resume screener with FAISS + Sentence Transformers for semantic skill matching across 9 categories. Includes 4 AI career tools: cover letter generator, ATS rewriter, skill gap advisor, and interview prep — powered by Groq LLaMA 3.3 70B.
Real-time financial news analyzer using Groq LLaMA, classifying market sentiment across 8+ articles with 75% accuracy. Built a RAG-powered chatbot extracting 20+ stock tickers with NLP and context-aware Q&A. Deployed on Render with sub-500ms response time.
Full-stack SaaS dashboard with JWT auth and RBAC across 4 roles — org-scoped data isolation on all routes. Real-time analytics with Recharts tracking sessions, revenue, and active users from MongoDB Atlas.
AI-powered full-stack app that generates accurate SQL from plain English using LLaMA 3.3 70B, eliminating the need for SQL knowledge. Supports multi-user SQLite and CSV uploads with dynamic schema extraction and live query execution.
Flask web app for multi-file PDF/DOCX Q&A and article summarization using Groq LLaMA 3. Features session history, PDF export, and drag-and-drop upload with smart context extraction.
AI-powered travel planner that generates personalized itineraries, suggests destinations, accommodations and activities based on user preferences using LLM reasoning.
AI-powered CRM lead importer that maps any CSV layout — Facebook Lead Ads, Google Ads, spreadsheets, other CRMs — into a fixed schema without assuming fixed column names. Groq-powered field mapping with enum validation, batch-level fault isolation, and API key rotation for quota resilience.
// 05 — Contact
Whether it's an internship, collaboration, freelance project, or just a chat about AI — I'm always open. Reach out via the form or find me on:
Message sent! I'll get back to you soon.
Or email me directly at harshupadhyay1414@gmail.com