Home
Team Name: Nexa
Team Logo:
Theme Music: "Awake" — Tycho. This was filled for grading purposes. Then Mona changed groups, so team can decide on different music.
Nexa is a mobile-first web app that simplifies reporting neighborhood issues to local government. Users submit a photo, voice note, or text description of a problem (e.g., a pothole, illegal dumping, vehicle emissions), and the system uses AI to classify the issue, determine the responsible agency based on GPS and jurisdictional data, and surface the official city 311 / Report-an-Issue page. Reports are stored in a personal dashboard with status tracking so users can follow up without starting from scratch.
The production app is deployed on Vercel and backed by Neon Postgres. The following features are implemented and working end-to-end:
| Feature | Description |
|---|---|
| Report wizard | Three-step flow: Describe → Review → Confirmed. Supports text, photo upload (drag & drop or camera), and GPS location detection. |
| Multi-LLM AI classification | Three providers (OpenAI GPT-4o-mini, Anthropic Claude 3.5 Haiku, Google Gemini 2.0 Flash) classify in parallel. A consensus engine picks the best answer. The review step shows a comparison panel with per-model latency and confidence. |
| Address autocomplete | Google Places suggestions as you type (falls back to Nominatim when no API key). Picking a suggestion sets both the address and GPS coordinates. |
| Official city form lookup | After classification, the review step shows the official city 311 / Report-an-Issue page for the user's location. Nexa never sends data to the external site. |
| Dashboard | Personal report history with status pills, category labels, timestamps, and two-step report deletion (owner-only, server-side auth). |
| Auth | Register, login, logout with JWT sessions. Session-aware navbar. |
| PostHog telemetry | Passive event tracking (report_classified, report_submitted) and session replays for measuring KPIs. |
Tech stack: Next.js 16 · React 19 · TypeScript · Tailwind CSS v4 · shadcn/ui · Prisma ORM · PostgreSQL (Neon) · Vercel · GitHub Actions CI
- Midpoint User Testing Plan
- Measure for Success (OKRs / KPIs)
- Customer Discovery Summary
- PRD (Google Doc)
 Keith Tawe Vhurumuku |
 Elias Chikwanda |
 Neil Shen |
David Lu
|
 Sarah Hashash |
| Member | Skills / Expertise | Personal Traits | Desired Growth | Gaps / Weaknesses | Email (@stanford.edu) |
|---|---|---|---|---|---|
| Keith Tawe Vhurumuku | Embedded Systems, FPGA/Digital Design, Computer Architecture, Full-Stack Web (Next.js/TypeScript), Verilog, C, Python, iOS Dev | Hardworking, Self-directed, Teaching-oriented | Intersection of hardware design and intelligent systems | AR/VR, UI/UX Design | tawekith@stanford.edu |
| Neil Shen | Front-end, product management | Team-oriented, hardworking | Working together to build an interesting project | Technical skills | neilshen@stanford.edu |
| Elias Chikwanda | Full Stack Web (Spring Boot, Angular, PostgreSQL, Docker) | detail-oriented, curious, collaborative | Distributed systems, scalable cloud architecture, and deeper ML/systems integration | Limited experience with large-scale production deployment and frontend design | eliaschi@stanford.edu |
| David Lu | Full Stack (react js, postgresql, nosql) | Curious, Hard working, Collaborative | Need finding high quality product and building for value | limited experience with building production systems | davidlu7@stanford.edu |
| Sarah Hashash | Web Development, AI/ML | Self-driven, Hardworking, Collaborative | Building useful products and learning more about full-stack development | Frontend design, large-scale deployment | shashash@stanford.edu |
- Embedded Systems & Hardware (bare-metal C, STM32, Raspberry Pi)
- FPGA / Digital Design (Verilog, pipelined CPU design)
- Computer Architecture
- Full-Stack Web Development (Next.js, TypeScript, React)
- iOS / Mobile Development
- Teaching & Technical Communication
- UI/UX Design
- GIS / Geospatial Data experience
- DevOps
- Slack: Insert Slack link here
- Email group: Insert team email or contact method here
- Repository: spr26-Team-24
- iMessage: Group chat