Full Stack AI

Qora Credit

SaaS product built to reduce manual work in financial operations. As a Full Stack Engineer, I led backend, applied AI, and product flows built for operations.

Role: Full Stack Engineer across backend, applied AI, and product

Problem: Automate financial data capture and analysis without depending on slow, fragmented manual processes.

Solution: Built an event-driven backend with multi-provider AI, voice transcription, and a frontend tailored for operations.

Impact: Reduced operational friction in financial capture and analysis.

Highlights

Designed an async BullMQ pipeline with deduplication for traceable processing.
Integrated Gemini, Claude, and OpenAI through LangChain by use case.
Added real-time voice transcription with Deepgram to reduce capture friction.
Implemented real-time notifications with Pusher for better visibility.
Secured authentication with Google OAuth2 + JWT in HTTP-only cookies.

Core stack

NestJS 11 Next.js 15 MongoDB BullMQ LangChain Deepgram SDK Pusher