Final-year B.Tech Computer Science student in Delhi, India. I build Python tools for scraping, data pipelines, and AI applications — and I put the working code where you can read it.
AWS Certified Cloud Practitioner · Python · pandas · scikit-learn · LightGBM · Playwright · SQL
web-scraper-demo — a scraper built the way one you depend on should be
Retry logic with exponential backoff, polite rate limiting, a Playwright fallback for JavaScript-rendered sites, and CSV/Excel/JSON export. Parsers are pure functions, so they test offline without hitting the network.
rag-chatbot-demo — document Q&A that shows its sources
Upload PDFs, ask questions, get answers with file and page citations. Embeddings run locally, so the whole thing works with no API key. Ask it something the documents don't cover and it says so instead of inventing an answer.
data-cleaning-demo — messy CSV to clean dataset, with an audit trail
Duplicates, six date formats in one column, currency symbols inside numeric fields, categories spelled twelve ways. Every change is logged, so when the row count drops you can see exactly where it went.
mcp-server-demo — a real MCP server for Claude
Read-only SQL, sandboxed file search, and a weather tool. Read-only is enforced by the database engine, not just by my code. 40 end-to-end tests over real stdio transport, including SQL-injection and path-traversal attempts.
Everything above is tested, and every repo has a test suite you can run yourself. Nothing ships that I have not run.
- Fiverr: My Fiverr profile
- Upwork: My Upwork profile
- Email: harshgupta93198@gmail.com
Open to Python automation, web scraping, RAG/LLM applications, Claude Code and MCP setup, and data cleaning work.