profile-image

Amdad Shabbir

Full Stack AI Engineer

Bangladesh

Language / Framework:

TypeScript, Python, Node.js, React.js, Next.js, React Native (Expo), FastAPI, NestJS, ExpressJS, LangChain, LangGraph

AI / LLM Engineering:

Prompt Engineering, LLM Function Calling, AI Agent Workflows, RAG Pipelines, Embeddings & Semantic Search, AI Observability, AI Cost Optimization, Prompt Versioning, Model Fallback Strategy

Databases / Backend Infrastructure:

MySQL, PostgreSQL, MongoDB, Redis, Pinecone (Vector DB), Neo4j (Graph DB), Supabase, Firebase, Prisma, Mongoose, Sequelize, Background Workers, Queue Systems

Deployment:

Git, GitHub, GitHub Actions, CI/CD, Docker, Linux, Turborepo, Network Security, Dokploy, Cpanel

AI Tools & APIs:

Claude, Cursor, Windsurf, Whisper (Speech-to-Text), ElevenLabs (Text-to-Speech), OpenAI API

Dev Tools & Libraries:

WebSockets / Realtime Updates, Audio Processing, HTML, CSS, Tailwind CSS, Material UI, Shadcn UI, Bootstrap, Redux Toolkit (RTK), Zustand, Formik, Zod, Axios, JWT, Framer Motion, Jest, Figma, Jira, Notion

Works

product-image

Chattrn

chattrn.com

Ai Agent for Customer Service

  • Contributed across the Next.js frontend, backend APIs, and agent/LLM flows so support teams get dependable AI-assisted tooling in production.
  • Worked end-to-end on typed UI, server endpoints, and AI-facing behavior—tools, validation, and realtime channels tied to the customer-service agent.
  • Helped wire WebSocket-driven experiences, conversation state on the server, and the integration surface between models and what users see in the app.
product-image

Talk Expense

talkexpense.com

Ai Agent for Personal Finance Management

  • Owned client, backend, and the voice expense stack end-to-end—including ASR in, LLM/tool orchestration, and TTS for spoken agent replies.
  • Shipped data models, auth, APIs, and LLM-backed categorization so outputs land as structured expense records.
  • Wired capture, agent logic, and synthesized speech in the UI: utterance → reply → persisted transaction.
product-image

Pocket School

pocketschool.academy

Platform for SWE Fundamentals

  • Contributed on the frontend only—shipping course flows, lesson UI, and state-heavy screens in React with Tailwind and Redux.
  • Implemented React Hook Form + Zod flows, accessible layouts, and responsive views matched to Figma while coordinating with the team’s API contracts.
  • Focused my work in the shared web client—reviews, refactors, and features stayed on the FE layer while teammates owned services and data.