# Sujal Goel I build AI and automation into real products and ship them live. - **Role**: Full stack developer · AI and automation - **Location**: Based in India, works worldwide - **Email**: hello@hecodesforme.com - **Site**: https://hecodesforme.com Top Rated full stack developer and AI automation engineer. Production Next.js, Python, scraping and data pipelines. Based in India, works worldwide. ## Projects ### Outer Banks Sharp Stays Vacation rental booking platform — https://sharp-stays.hecodesforme.com _live · used daily_ Built from scratch for a client with 11 properties. It takes real money every day. Stripe checkout with the full price breakdown, nightly rate, cleaning, pet fee and taxes. Two-way iCal sync against Airbnb feeds so availability stays honest. Magic-link login, Twilio SMS for ops, and one admin calendar showing all 11 properties at once. ### Carolina Connections Academy Real estate license school — https://carolinaconnectionsacademy.com _live · NCREC provider #1853_ The site an NCREC-approved school runs its enrolments on. Course catalogue, session scheduling, student login and tutoring, all under one brand. It has to hold up for people spending real money on a licence. ### ProTrain Athlete Speed training program — https://protrainathlete.com _live_ A conversion site for a coach who guarantees the result or refunds it. Built around one promise and one action. Everything on the page either supports the guarantee or gets out of the way of the booking. ### The Builder Plan Quiz funnel — https://thebuilderplan.com _live · 4-step qualifier_ A four-question funnel that qualifies people before anyone picks up a phone. Insurance is a slow, high-friction sell. The quiz does the filtering up front so the conversation starts with someone who already fits. ### Scraping infrastructure Automation at scale _6 months · 0 bans_ Ran for months on heavily protected sites without a single block. The hard part was never the parsing. It was staying boring enough to be ignored: pacing, rotation, and backing off before anything noticed a pattern. ### Lead intelligence pipeline Data + AI _1.19M → 173k_ Turned 1.19M raw rows into 173k leads worth calling. Scores about 196k records a run through OpenAI with retries and a hard cost ceiling, then loads straight into Postgres. Most of the work was throwing rows away correctly. ### Identity sync platform Infrastructure _drift surfaces itself_ Pulls identity from JumpCloud and Entra and makes disagreement visible. Normalizes both into one schema and runs deterministic diffs, so drift shows up on its own instead of waiting for someone to notice. Strict Pydantic validation, retries, pytest coverage. ### RAG fact-checker AI verification _claim-by-claim_ Takes a video, pulls out every claim, checks each one against live sources. The interesting failure was never a wrong verdict. It was a claim the extractor never saw. ## Links - Upwork: https://www.upwork.com/freelancers/~0137f5d7bf36c18c36 - GitHub: https://github.com/sujalgoel - LinkedIn: https://www.linkedin.com/in/sujalgoel - X: https://x.com/hecodesforme - Instagram: https://www.instagram.com/friendlyfatkid/