Enterprise automotive marketplace

Inherited Architecture Challenges

We took over the project after 10 months of prior development, inheriting significant technical instability:

• Architectural Instability: The codebase lacked structure, with single files reaching 30,000 lines of code, widespread duplication, and critical TypeScript errors throughout.
• Inconsistent AI-Generated Code: Unreviewed AI-assisted code segments had introduced logic inconsistencies, making the system difficult to maintain or extend reliably.
• SEO Architecture Gap: The original SPA-only architecture prevented indexing of dynamic vehicle listings — a core business requirement.
• Infrastructure Gaps: No CI/CD pipeline, compromised database integrity, malfunctioning fintech integrations, and no mobile-responsive design — all against an imminent launch deadline.

AI Engineering Governance

We established structured controls to make AI-assisted development predictable and maintainable:

• Engineering Standards: Implemented Claude Code with strict architectural rules, custom skills, and automated code review workflows to ensure consistent output quality.
• Codebase Stabilization: Removed dead code and restructured the system architecture, restoring predictability and measurable performance.

SEO Architecture via Next.js

The frontend was fully rebuilt to meet search engine optimization requirements:

• SSR/SSG Strategy: Migrating to Next.js made every brand, model, and derivative instantly indexable.
• Dynamic Content Routing: A URL hierarchy that automatically generates SEO-optimized landing pages based on live vehicle inventory.

Fintech & Marketplace Features

High-complexity features enabling a end-to-end purchase journey:

• Seriti API Integration: Credit calculators, insurance modules, and automated financing applications.
• Typesense Search: Fast, AI-assisted search with "similar vehicle" recommendations and multi-parameter filtering.
• AI Content Pipeline: Automated image processing for background removal and interior/exterior classification to maintain consistent visual presentation across inventory.

DMS & Back-Office

The internal Dealer Management System was re-engineered for operational scale:

• Stock & Lead Management: Real-time data validation for stock imports, test-drive scheduling, and lead generation.
• Granular RBAC: A hierarchical Role-Based Access Control system for dealers, brand managers, and content administrators.

Timeline & Risk Management

When Sigmalion engaged, 6 months remained until the final deadline. The client's original plan allocated 5 months for development and 1 month for finalization.

• Accelerated Delivery: Leveraging AI-governance workflows, we completed the deep refactoring and stabilization phase in 3 months.
• Risk Mitigation: Delivering the core system approximately 2 months ahead of the revised schedule eliminated time pressure and provided a meaningful buffer for thorough testing and a stable production launch.

The Impact

Sigmalion delivered measurable engineering improvements, restoring the platform from a critically unstable state to approximately 93% production readiness — based on client acceptance criteria — within the agreed timeline.

Beyond the product itself, we provided a transparent, scalable business asset: a healthy database, professional AWS CI/CD infrastructure, and a solid SEO foundation. The project was handed to the client's maintenance team in a well-documented, stable condition.