These aren’t hypotheticals. These are real patterns we’ve seen across hackathons on Reskilll – projects that caught a company’s attention and led to direct offers.
Story 1: The AI Triage Bot That Got a HealthTech Offer
What they built: A WhatsApp bot that asks patients symptoms in Hindi, categorizes urgency (emergency / urgent / routine), and helps them find the right department.
Tech: Python + WhatsApp Business API + Claude API for symptom understanding + multilingual support
What happened: A HealthTech startup was a hackathon mentor. They saw the team’s approach to handling edge cases (what if the patient describes something ambiguous?) and offered the backend developer an internship that converted to full-time in 3 months.
Why it worked: The team didn’t just build a chatbot – they thought about the REAL user (rural patient, basic phone, speaks Hindi). That product thinking impressed the startup more than the code.
Story 2: The MCP Server That Impressed an AI Startup
What they built: An MCP server that connected Claude to a college’s exam database. Students could ask “When is my algorithms exam?” and get instant answers.
Tech: TypeScript + MCP SDK + SQLite + semantic search
What happened: The project got shared on Twitter. An AI tooling startup saw it, reached out, and hired the builder to work on their MCP integrations. Zero traditional interview – they just reviewed the GitHub repo.
Why it worked: MCP is the hottest protocol in AI. Building a working MCP server showed the developer was ahead of the curve.
Story 3: The Agricultural Dashboard That Won a Government Contract
What they built: A dashboard combining mandi prices + weather + satellite data to help farmers decide the optimal day to sell their crop.
Tech: React + Python ML model + Government mandi API + weather API + maps
What happened: The team won their SIH grand finale. The government ministry that sponsored the problem statement asked if the team could develop it into a production system. The team lead got a contract to build it.
Why it worked: They solved the EXACT problem the government described, with real data, and a working demo. The ministry didn’t need to search for vendors – the solution was already built.
Story 4: The Resume Ranker That Got the Team Into YC Interviews
What they built: An AI that ranks job applications against a role description, explains WHY each candidate is a match or mismatch, and suggests interview questions specific to each resume.
Tech: Next.js + Claude API + PDF parsing + scoring algorithm
What happened: They built it at a hackathon, polished it for 2 more weeks, then applied to Y Combinator with it. Got an interview. Didn’t get in, but the experience and YC interview on their resume got both founders hired at top startups.
Why it worked: They identified a problem every company has (resume screening is painful) and built something a non-technical HR person could use. Business viability was obvious.
Story 5: The Code Review Agent That Landed a DevTools Job
What they built: A GitHub bot that automatically reviews PRs using Claude – checks for security issues, suggests improvements, and explains complex changes to reviewers.
Tech: GitHub webhooks + Claude API + PR commenting API + diff parsing
What happened: They open-sourced it on GitHub. It got 200 stars. A DevTools company saw it, loved the approach, and hired the developer to build similar features in their product.
Why it worked: They made it public, documented it well, and it solved a real developer pain point. The GitHub repo was their resume.
The Pattern
Every story follows the same formula:
- Built something that solves a REAL problem (not a demo)
- Made it work live (deployed, demo-able)
- Made it visible (GitHub, Twitter, hackathon stage)
- Someone who could hire them SAW it
The hackathon gives you steps 1-3. Step 4 happens naturally when you build in public.