A real estate CRM built solo on weekends for a private client in Bangalore — and delivered. I own the entire product: the backend API, a cross-platform mobile app for field agents, and a web portal for admins. It’s actively used in production today.

What I Built

• Mobile App for Field Agents — One codebase, runs on iOS, Android, Web, and PWA (Expo + React Native). Agents manage their leads, log calls, set follow-up reminders, and track daily performance — all from their phone.
• Web Admin Portal — Admins manage property listings, import leads from Excel, monitor team performance, and view conversion analytics. Deployed on Cloudflare Pages.
• Lead Pipeline — Buyer inquiries move through a full sales funnel: new → contacted → interested → negotiation → converted/lost. Every action is permanently logged — nothing is ever overwritten, full history always available.
• Role-Based Access — Admins see all leads and agents. Agents see only what’s assigned to them. Enforced at the API level.
• Property Catalog — Residential and commercial listings with image galleries, per-unit pricing (2BHK, 3BHK, etc.), and delivery date tracking.
• Bulk Lead Import — Upload an Excel sheet and the system ingests it automatically with throttling to avoid overloading the database.
• Agent Location Tracking — GPS check-ins stored in a partitioned database table, opt-out per agent.
• Backend — Async Python API (FastAPI + SQLAlchemy), JWT auth, deployed on DigitalOcean.

rs-1

rs-4

rs-5

rs-6

rs-7

rs-8

rs-9

rs-10

rs-11

rs-12

rs-13

rs-35

rs-36

rs-14

rs-15

rs-16

rs-17

rs-18

rs-19

rs-21

rs-22

rs-23

rs-27

rs-28

rs-29

rs-30

rs-31

rs-32

rs-33

rs-34

rs-37

rs-38

rs-39

Stack: Python/FastAPI · PostgreSQL · Redis · React Native (Expo) · React 19 · TailwindCSS · DigitalOcean · Cloudflare Pages · S3

Rosche is a full-stack, enterprise-grade Risk and Vulnerability Management Platform designed to help security teams manage the full lifecycle of security engagements — from asset discovery and vulnerability assessment to remediation tracking and executive reporting.

Phase 1 — Foundation — NestJS backend scaffold, PostgreSQL schema, JWT auth, base CRUD for assets and engagements.

Phase 2 — Core Features — Vulnerability management, CVSS/DREAD scoring engine, RBAC and permissions system, React frontend with Material UI.

Phase 3 — Advanced Features — AI-powered CVSS generation (Google Gemini), PDF report generation, MeiliSearch full-text search across findings, assets, and engagements, Redis caching.

Phase 4 — Infrastructure — Docker/Nginx containerization, AWS ECS Fargate deployment, CloudFormation IaC, AWS Cognito with MFA.

Phase 5 — Polish & Testing — Playwright E2E tests, Storybook component documentation, admin panel (Refine), multi-tenancy hardening.

Architecture — 7-service Docker Compose setup: NestJS API · Admin API · React frontend · Admin panel (Refine) · PostgreSQL · MeiliSearch · Nginx reverse proxy. Cloud deployment via AWS ECS Fargate with ALB, KMS-encrypted secrets, and CloudFormation IaC.

Backend: NestJS · TypeScript · PostgreSQL (TypeORM) · Redis · MeiliSearch · AWS (ECS, S3, KMS, Cognito, SES)

Frontend: React 19 · Vite · Material UI v7 · Redux Toolkit · CKEditor 5 · ECharts · @react-pdf/renderer

I contributed to the development KhetAI’s MVP, an android application designed to facilitate precision farming for farmers through machine learning predictions. My role involved seamlessly integrating Google Maps into the application, enabling precise mapping of farm boundaries.

I worked on the following tech stack:

• Android Studio
• Java, Kotlin
• Firebase