AI Engineer Projects in RexAI Engineer Projects in RexVenture Vault Capital - Base44 to Supabase Migration
Migrated a funding platform from Base44 to Supabase, modernizing its backend architecture while preserving existing application functionality. The project covered PostgreSQL database migration, Supabase Authentication, data modeling, access control/RLS, API integration, codebase updates, testing, and deployment validation.
Key Highlights:
✔ Base44 → Supabase/PostgreSQL migration
✔ Supabase Auth & user management
✔ Database schema and data integrity migration
✔ Row Level Security (RLS) and permissions
✔ Backend/API compatibility updates
✔ Testing, deployment, and migration documentation
Stack: Supabase, PostgreSQL, GitHub, REST APIs, Authentication, RLS NJBSoft Compliance Software Management (secure compliance platform for inspections, permits, documents, and public-sector workflows)
NJBSoft is a compliance management platform used to unify inspection data, permit records, operational workflows, and documentation in one secure system. I helped build the core web application and backend workflows around document access, inspection records, reporting, and role-based permissions. The important overlap here is tenant-safe compliance data: every API request needed clear organization/project context, so one client’s records could not leak into another client’s workspace. 𝐄𝐧𝐭𝐞𝐫𝐚 - 𝐑𝐞𝐬𝐢𝐝𝐞𝐧𝐭𝐢𝐚𝐥 𝐑𝐞𝐚𝐥 𝐄𝐬𝐭𝐚𝐭𝐞 𝐈𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭 𝐏𝐥𝐚𝐭𝐟𝐨𝐫𝐦 (𝐀𝐈-𝐩𝐨𝐰𝐞𝐫𝐞𝐝 𝐩𝐫𝐨𝐩𝐞𝐫𝐭𝐲 𝐚𝐜𝐪𝐮𝐢𝐬𝐢𝐭𝐢𝐨𝐧 𝐚𝐭 𝐬𝐜𝐚𝐥𝐞)
◾ Frontend: React, Next.js, TypeScript, Tailwind CSS, Recharts, Mapbox GL JS
◾ Backend: Python, FastAPI, Node.js, PostgreSQL, Redis, Celery, REST & GraphQL APIs, OpenAI API, custom ML pipelines
◾ Infra: AWS (EC2, RDS, S3, Lambda), Docker, Kubernetes, GitHub Actions CI/CD, Elasticsearch
Entera is a platform built for institutional investors to discover, underwrite, and acquire single-family homes at scale - think hundreds of offers a day across multiple markets. I worked on the AI-driven deal analysis engine that pulls in MLS data, rental comps, and neighborhood signals to surface investment-grade properties ranked by projected yield and risk. 𝐅𝐫𝐚𝐦𝐢𝐪 (𝐒𝐚𝐚𝐒 𝐦𝐨𝐜𝐤𝐮𝐩 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐨𝐫 𝐰𝐢𝐭𝐡 𝐀𝐈 𝐜𝐨𝐧𝐭𝐞𝐧𝐭 𝐩𝐥𝐚𝐧𝐧𝐢𝐧𝐠 & 𝐚𝐮𝐭𝐨 𝐬𝐨𝐜𝐢𝐚𝐥 𝐩𝐨𝐬𝐭𝐢𝐧𝐠)
◾ Frontend: React 19, TypeScript, MUI component library, GraphQL mutations & subscriptions (AWS AppSync), mobile-first layouts
◾ Backend: Python 3 Lambdas, OpenAI API for content planning, image composition pipeline, Instagram/Facebook/X third-party API integrations, DynamoDB
◾ Infra: AWS Amplify Gen 2 (TypeScript IaC), Step Functions, S3 + CloudFront with Lambda@Edge image optimization, Cognito multi-tenant auth, X-Ray tracing
I owned the entire stack: the React 19 frontend including the real-time workflow status UI, the Step Functions state machine that orchestrates scraping → AI content planning → image generation → social publishing in sequence and parallel, and the CloudFront + Lambda@Edge layer that serves optimized assets at the right dimensions per platform. 
