GigShield: Parametric Income Insurance for Gig Workers

Delivery riders in India are paid weekly and live on thin cash buffers. One day of heavy rain, a hazardous AQI reading or a platform outage can wipe out a day or two of income, and traditional insurance, with its paperwork, adjusters and reimbursement weeks later, doesn't fit a loss that small and that sudden. GigShield was our answer: a parametric income-protection platform built by a team of three, with Rishabh and Abhinav.

Parametric cover pays a fixed amount when an objective, external signal crosses a threshold, rather than reimbursing a loss after someone proves it. We priced it the way riders think about money, weekly: Basic at 79 rupees, Standard at 129 and Pro at 179, each with its own daily payout cap and covered hours. The product deliberately covers lost earning hours only, not health, accidents or vehicles.

My main piece was the trigger engine. It covers seven kinds of disruption, including heavy rain, heatwaves, AQI spikes, platform outages, curfews, local strikes and zone closures, and rather than a bare yes or no, each trigger gets a confidence score. That score starts from a base value per trigger type, rises with how reliable the weather data was, and falls for riders with a higher risk profile, ending up labelled Strong, Normal or Weak. Weather comes from Open-Meteo with a 2.5-second timeout and an automatic fallback to wttr.in, so one flaky API can't quietly stop the engine. Cooldown and de-duplication windows stop a single storm from firing again and again, and every trigger is written to an audit trail that syncs to the backend.

I also built the payout side. A claim moves through a visible lifecycle, pending verification, verified, processing and settled, and riders get a downloadable receipt generated with jsPDF, with their payout history loaded from the backend. In this build, payouts are simulated; there's no real payment rail connected.

For fraud, Rishabh and I built a liveness check on risky sessions. The app asks for a random hand gesture, and MediaPipe's hand-landmark model classifies it in the browser from the positions of the fingertips. I built the admin operations and fraud-review screens around it, plus role-based route protection so riders and admins only reach their own views. Riders also get a weather radar map built with Leaflet, and the landing page uses React Three Fiber for its 3D scene. Rishabh and Abhinav built most of the authentication, the Express server and the Groq-powered help chat.

Looking back, two things stand out. First, the cooldown and de-duplication state lives in the browser's storage, which is fine for a demo but belongs on the server in anything real. Second, our trigger thresholds are reasonable guesses; calibrating them properly would need historical weather and earnings data we didn't have. What I'd keep is the core idea: for someone paid weekly, the most useful insurance is the kind that pays without being asked.