e-Nivaran

Citizen complaint systems are often fragmented across departments, slow to route urgent cases, and opaque after submission. Residents need one place to report issues, understand their status, and surface problems that demand faster attention.

I led the product and full-stack architecture from idea to public release. I built the complaint workflow, connected the classification service, designed severity-based routing and community upvoting, and prepared the platform for its Government of Delhi showcase.

A Next.js 15 application handles the citizen experience and operational views. FastAPI serves the complaint-classification layer, while MongoDB Atlas stores complaint state. Severity scores and community upvotes contribute to escalation so urgent, widely felt issues become visible sooner.

• Separate product and classification services

  Keeping the FastAPI classifier independent from the Next.js application made the ML workflow easier to iterate without coupling it to the citizen-facing release cycle.

• Combine severity with community signal

  Classification alone can miss local impact, so escalation also considers upvotes. The trade-off is added moderation responsibility, but the result better reflects what residents experience.

• Make status visible

  Complaint tracking was treated as a core product feature rather than an administrative detail, giving citizens a clearer view of routing and progress after submission.

e-Nivaran reached the Hack4Delhi Grand Finale at Bharat Mandapam, placing among the top 30 of more than 1,500 teams. The public platform received 1,500+ visits and was demonstrated to Government of Delhi representatives.

“Civic technology earns trust through visible follow-through; intelligent routing matters only when citizens can also see what happens next.”