PermitAI (Phase 1 Architecture)
PropTech regulatory database and state machine platform for municipal permit workflow automation
Problem
Real estate developers and construction firms face crippling project delays when navigating fragmented municipal and government permit approval pipelines. Building applications must pass through dozens of disjointed regulatory bodies (environmental, structural, zoning, and fire safety), with zero centralized tracking or document verification transparency—turning permit acquisition into a multi-month bureaucratic black box.
Solution
Architected and developed Phase 1 of an advanced digital permit management and compliance verification platform. Centralizes regulatory documentation, verifies submission completeness via automated document pipelines, and orchestrates complex, multi-agency permit applications through a deterministic, auditable workflow state machine.
How It Was Built (Architecture & Implementation)
Multi-Agency State Machine Engine (Node.js & PostgreSQL)
Designed a highly scalable, deterministic workflow state engine utilizing PostgreSQL JSONB schemas and state transition tables. Every permit application tracks precise approval stages across independent regulatory agencies, preventing out-of-order approvals or skipped compliance checks.
Document Verification & Compliance Pipeline
Built modular document ingestion APIs that validate uploaded architectural drawings, land titles, and structural certifications against required regulatory checklists before applications can enter agency queue tracks.
Audit-Ready Regulatory Database Architecture
Engineered immutable audit logging tables ensuring every agency review, comment, approval signature, or document rejection is permanently timestamped and traceable for legal and municipal compliance.
Developer & Agency Portal APIs (REST)
Developed clean RESTful API contracts enabling both applicant-facing dashboards and agency-facing review portals to interact with the core workflow engine seamlessly.
How I Improved It Over Time (Evolution & Iteration)
[ Phase 1.0 (Core Database & State Machine Architecture) ]
Designed and implemented the core relational schema, state transition engine, and API endpoints, establishing a rock-solid foundation for multi-stage government approvals.
[ Phase 1.5 (Document Pipeline & Audit Logging Validation) ]
Completed the document validation gateway and immutable audit logging layers, proving that complex municipal workflows can be modeled cleanly in code without state leaks.
[ Phase 2 Roadmap (AI Document Parsing & Live Agency Integration) ]
Preparing architecture for Phase 2 validation, which will introduce automated OCR/LLM extraction of architectural drawing metadata and direct API hooks into municipal databases.
Outcome & Business Impact
Phase 1 core backend architecture and state machine engine completed successfully. Validated the technical feasibility of automating multi-agency permit lifecycles through deterministic software engineering. Currently in technical staging awaiting Phase 2 field metrics.
My Role & Execution
Core Backend & Systems Architect. Designed the relational and JSONB database schema, built the Node.js/Express REST APIs, authored the state machine transition logic, and implemented document verification pipelines.
Key Learnings
- Architecting deterministic, complex state machines for long-running, multi-stakeholder government workflows
- Designing flexible data models capable of handling evolving municipal compliance and document verification rules
- Structuring immutable audit trails and secure document handling pipelines for regulatory software