TriformaIQ Real Estate Studio

#python

#fastapi

#sqlmodel

#postgresql

#postgis

#pgvector

#alembic

#celery

#redis

#docker

#llms-(anthropic-api)

#rag

#stripe

#esri-arcgis

#auth0

#react

#typescript

#pytest

#playwright

#claude-code

Manuel Solalinde

October 14, 2026

Website: https://www.triformaiq.com/

Developer: Manuel Solalinde (manolosolalinde@gmail.com)

TriformaIQ Real Estate Studio is an AI-powered, multi-tenant SaaS platform for commercial real estate teams. It brings site selection, entitlements, due diligence and audit-ready documentation into one workspace, so that teams can evaluate locations with data instead of guesswork.

Overview

The platform combines interactive GIS with weighted multi-criteria scoring to compare candidate sites side by side. It also offers secure per-project document storage and an immutable audit trail. I work across the stack, mainly on the API and data layer, building AI features, SaaS infrastructure and geospatial capabilities for a product heading to launch.

Contributions

LLM-powered document analysis (RAG): document ingestion and chunking, vector embeddings and semantic search on PostgreSQL/pgvector, structured LLM outputs grounded in verifiable source citations, and a regression evaluation harness to track prompt quality over time.
SaaS billing and licensing: payment provider integration through webhooks, subscription lifecycle management, and plan-based feature gating enforced at the API layer.
Platform capabilities: role-based approval workflows, per-user and per-organization rate limiting of third-party APIs, and geospatial data features, all with strict tenant data isolation.
AI-assisted engineering: Claude Code workflows (custom skills and multi-agent orchestration) that automate ticket analysis, implementation, code review and end-to-end verification across the backend and frontend repositories.

Architecture

Backend: a FastAPI service organized as domain modules (API, application services, domain, repositories), with SQLModel/Alembic on PostgreSQL + PostGIS and Celery/Redis for background work.
Frontend: a React + TypeScript dashboard that consumes an auto-generated OpenAPI client, with ArcGIS-based mapping.
Multi-tenancy: every query is scoped to the authenticated tenant, which is derived from the JWT and never taken from client input.
Quality: unit, integration and API tests with Pytest, Playwright end-to-end tests, and containerized development with Docker Compose.