University TA Agent
Modular skill-registry / router / executor pattern with RAG-based Q&A, auto-grading, and Zod-validated tool schemas.
I design modular NestJS monorepos, event-driven microservices, and the multi-cloud infrastructure they run on — then ship production LLM/RAG and agentic AI pipelines on top. Based in Dhaka, Bangladesh 🇧🇩
I'm a Principal Engineer with 9+ years spanning NestJS/Node.js backend architecture and multi-cloud infrastructure. I build modular NestJS monorepos, async queue systems with BullMQ, and event-driven microservices — while owning the AWS/GCP infrastructure, Terraform IaC, and CI/CD pipelines they run on.
That full ownership of the delivery stack — TypeScript services, container orchestration with Docker & Kubernetes, and production LLM/RAG and agentic AI pipelines (LangChain, LangGraph, Claude & OpenAI) — lets me cut a problem at whatever layer it actually lives in.
The results speak in numbers: query load from 50s to under 2s, backend throughput up 70%, deployment time down 60%, data-loss risk down 80%, and 20+ AI agents in production.
Modular NestJS monorepos, typed providers, Zod validation, async BullMQ pipelines, and event-driven microservices.
AWS/GCP infrastructure, Terraform & CDK IaC, Kubernetes orchestration, CI/CD, and observability that stays green.
Model-agnostic RAG platforms, LangGraph agentic workflows, pgvector search, and Claude/OpenAI pipelines in production.
From backend developer to principal engineer — eight roles across Bangladesh, the US, and Canada.
Production systems where backend architecture, cloud infrastructure, and applied AI meet.
Modular skill-registry / router / executor pattern with RAG-based Q&A, auto-grading, and Zod-validated tool schemas.
NestJS worker on AWS ECS Fargate + SQS with Claude Batch API — multimodal RAG using Whisper transcription and Gemini embeddings.
GPT-4 research platform with MongoDB Atlas, PDF export, query classification, and external search routing.
Serverless microservice platform — full IaC-driven deployment on AWS Lambda with DynamoDB, Cognito, S3, SES, and SQS.
Multi-service AI assistant on GCP with Docker — LLMs, vector search, agentic tool integrations, and Cloud Run CI/CD.
Multi-tenant SaaS — NestJS API, Next.js frontend, Docker, Kubernetes, Nginx — full CI/CD and containerized deployment.
The stack I reach for across backend, infrastructure, and AI.
Open to principal and staff engineering roles in backend, cloud, and applied AI — remote or Dhaka-based. The fastest way to reach me is email.