B2B Scraping & Analytics Tool

A full-stack B2B data pipeline that scrapes structured and unstructured data from multiple sources concurrently, enriches it using LLMs for classification and validation, and surfaces the results through an analytics dashboard — with automated alerts and reports delivered via Discord and email.

Built a multi-threaded, concurrent scraping engine using Python's Requests library to pull data from multiple sources in parallel, significantly reducing collection time. Used Pandas for cleaning, deduplication, and transformation before persisting structured records to MySQL.

Integrated Groq and Gemini to classify, tag, and validate scraped records — filtering noise, standardizing inconsistent fields, and flagging anomalies that would otherwise require manual review. This kept the dataset clean without a dedicated data-entry step.

Built a FastAPI backend with a Node.js service layer for orchestration and scheduling, and a Next.js dashboard for visualizing scraped data and trends. Wired up Discord notifications for pipeline status/errors, Resend for email reports, and Web3Forms for lightweight external form intake — all deployed on AWS.