Five things
I actually built.

Not a booking app — an intelligent travel ecosystem. AI itineraries, budget planning, personalised recommendations, and a travel assistant that learns how you move.

• ◆AI-generated multi-day itineraries with budget guardrails
• ◆Modular architecture for future travel modules (visa, food, safety)
• ◆Flutter mobile client, Node backend, ML services

A government-focused AI emergency call assistant designed for Indian police. The long game: deploy nationally, cutting distress-call response times in half.

• ◆Voice-first NLP tuned for regional Indian languages
• ◆Auto location extraction & nearest-unit routing
• ◆Designed for pilot with a state police department

AI-driven medical emergency reporting for hospitals and first responders. Turns a chaotic call into a structured, prioritised intake before the ambulance arrives.

• ◆Severity classification with a lightweight ML model
• ◆Auto-notification to nearest hospitals
• ◆Pre-fills ER intake forms from the call

A quiet productivity app that pairs task management with lightweight budgeting. Built end-to-end as an exercise in shipping clarity, not features.

• ◆Offline-first with local persistence
• ◆Weekly budget rollovers with visual pacing
• ◆Zero-config onboarding

Started as coursework, ended as a real internal tool: role-based accounts, fine calculation, and a search that actually finds books.

• ◆Role-based auth
• ◆Search with fuzzy matching
• ◆Admin reporting