I'm a Backend and AI Engineer based in Dhaka, with about 2.5 years of building autonomous AI agents, LLM-powered workflows, and automation systems that run in production. A lot of my work lives at the intersection of voice agents, RAG and Graph RAG pipelines, and n8n automations, all sitting on solid backend engineering (Node.js, FastAPI) and DevOps (Docker, Kubernetes, CI/CD on AWS and Azure).
I currently work as a Backend and AI Engineer at AI Buddy Catalyst Labs, and I take on freelance automation work for teams in the US and Canada. I studied Computer Science at BRAC University (CGPA 3.93/4.00, Highest Distinction), where my research focused on how large language models hold knowledge in the Bangla language.
The short version: I build AI features that ship, automations that quietly save people hours, and backends that hold up as they grow.
| 🚀 2.5+ | 👥 10,000+ | ⚡ 60%+ | 🌍 US & CA |
|---|---|---|---|
| Years building AI systems | Users on a platform I built | Manual work cut with automation | International clients served |
| Area | What that looks like | |
|---|---|---|
| 🤖 | AI agents and agentic systems | Autonomous agents, Claude Skills, multi-agent orchestration, and tool calling |
| 📞 | Voice AI agents | Autonomous front-desk and sales agents with natural language understanding and real-time decisions |
| 💬 | Chatbots, RAG and Graph RAG | Retrieval pipelines with vector databases (Pinecone, Weaviate, Qdrant), shipped for edtech and Canadian clients |
| 🧠 | LLM routing and fine-tuning | Routing across OpenAI, Claude, and Gemini, plus fine-tuning with Unsloth and LLaMA-Factory |
| ⚙️ | n8n and workflow automation | Wiring tools together so the repetitive work runs on its own, cutting manual effort by more than half |
| 🧲 | Web scraping at scale | Selenium, proxies, and Cloudflare bypass, from raw scraping to clean, structured data |
| 💻 | Backend and full-stack | Node.js, Express, NestJS, FastAPI, and React, from idea to a deployed app |
| ☁️ | DevOps and cloud | Docker, Kubernetes, Terraform, and CI/CD on AWS and Azure |
| Project | What it does | Stack |
|---|---|---|
| 📞 Voice AI front-desk and sales agent | An autonomous calling agent that understands natural language, keeps context, and makes decisions in real time | Voice APIs LLM Python |
| 🔎 insta_rag (open source) | A production-grade RAG pipeline I built and documented for others to adopt | RAG Python Vector DB |
| 🐙 Octobrain EdTech platform | Backend and database for a student platform serving 10,000+ users, handling 10K+ student records | Node.js Express PostgreSQL Svelte Azure |
| 🕷️ Instagram scraper platform | A full-stack scraping system with async queues, serving multiple concurrent users on Azure CI/CD | FastAPI Node.js React RabbitMQ MongoDB |
AI, LLMs and automation
Backend and frontend
Data, cloud and DevOps
- AI Buddy Catalyst Labs, Backend and AI Engineer (Jun 2025 to present). Leading client projects end to end, building Graph RAG and RAG pipelines, routing multiple LLMs, and shipping microservices on AWS and Azure. Open-sourced
insta_rag. - Self-employed, AI Automation Engineer and Integration Specialist (Nov 2023 to present). Built an autonomous Voice AI agent, delivered n8n automations for US clients that cut manual work by 60%+, and built the Octobrain platform serving 10,000+ users.
- Kater & Out (USA fintech), Trainee Software Developer (Mar 2024 to Oct 2024). Helped build a catering platform MVP with Express, React, and Tailwind.
- BRAC University, Teaching Assistant for Data Structures and Programming (Nov 2024 to May 2025). Mentored 300+ undergraduates.
- 🥈 First Runner-Up and Global Nominee at the NASA International Space Apps Challenge 2022
- 🏅 Finalist at the Bangladesh AI Hackathon 2025
- 🎓 CGPA 3.93/4.00, Highest Distinction, 9x Vice Chancellor's List, 50% merit scholarship at BRAC University
- ☁️ Approver at the Cloud Native Computing Foundation (CNCF)
- 🎓 Microsoft Learn Student Ambassador (2022 to 2023)
- 📄 Co-author of "Understanding QA Generation: Extracting Parametric and Contextual Knowledge with CQA for Low-Resource Bangla Language" (2025), submitted to the ACL Rolling Review
