building useframeAI try!
I build AI products end to end - RAG pipelines, multi-agent systems, the UI people actually touch. idea to shipped, owning the full stack and the outcome
Experience
Tecnoprism
Full-Stack AI 2025 – Present
- Built and deployed a RAG pipeline using LangChain, integrating multiple LLM APIs with custom evals for output quality - serving 100K+ monthly requests and 8-10K+ concurrent users with real-time SSE sync.
- Engineered 15+ event-driven Node.js microservices using RabbitMQ and BullMQ, containerized with Docker and deployed on Kubernetes via Helm.
Alois Solution
Full-Stack 2023 - 2025
- Built a Next.js application from scratch with a PostgreSQL backend, implementing virtual scrolling to render large datasets smoothly, cutting render time by ~40%.
- Owned the CI/CD pipeline (GitHub Actions) with Nginx as reverse proxy for frontend deployments, eliminating manual steps and achieving zero-downtime releases.
AI Products
Full-stack AI chat app with conversation history, user profiles, Razorpay payments, and built-in CRM.
Next.js MongoDB Redis Express.js Razorpay AWS
Real-time messaging platform with multi-model AI for abuse detection and interest-based ad targeting.
Next.js Express.js Mongoose Stripe WebSocket DeepSeek MongoDB
A multi-agent SaaS platform that transforms startup ideas into structured business MVP applications using multi-model orchestration.
Next.js OpenAI PostgreSQL Docker Drizzle K8S Hybrid-RAG Node.js
open source
twentyhq/twenty
Open-source CRM · Node.js / TypeScript
- Fixed edge-case null handling in the metadata engine for custom field resolution
- Improved API response typing for relation fields in the GraphQL layer
novuhq/novu
Notification infrastructure · MongoDB / Next.js
- Contributed MongoDB index optimization for high-volume notification queries
- Patched Next.js subscriber portal to correctly handle missing provider configs
TypeScript ✱ Node.js ✱ React ✱ Next.js ✱ Python ✱ Tailwind ✱ MongoDB ✱ Pinecone ✱ PostgreSQL ✱ Redis ✱ RAG ✱ Applied-AI ✱ Docker ✱ K8S ✱ Vim ✱