portfolio

Work

Three full-stack AI systems — detection, retrieval, and multilingual support — each with deterministic security boundaries and comprehensive test coverage.

selected systems

01/Averqen

v1.0.0

AI-assisted security investigation with deterministic security boundaries.

Deterministic detection · Grounded AI · FORCE RLS · Controlled response

1,493 tests · 27 FORCE-RLS tables · 58 OpenAPI paths · AI trust boundary

View engineering case study

02/AtlasCore

UI v2

Secure enterprise AI infrastructure for knowledge, retrieval, and grounded AI workflows.

FORCE RLS · Hybrid retrieval · Grounded AI · Workspace UI

717 backend tests · 46/46 evals · FORCE RLS · pgvector · UI v2

AtlasCore workspace dashboard with sidebar navigation

View engineering case study

03/VaaniDesk

v1.0.1

Multilingual AI Customer Support Platform

Agents · RAG · Security · Evaluation

206 backend tests · 113 evals · 40 security tests · 14 E2E

View engineering case study