AI Agent Developer & Researcher | RAG & LLM Specialist | Full-Stack Developer
Specializing in AI agent and chatbot development, LLM-driven agentic systems, and retrieval-augmented generation (RAG) architectures. Developed by the BanglaCodeAct Project — recognized for pioneering work on Bangla-to-Python code generation using multilingual AI agents at AACL-IJCNLP. Creating intelligent solutions powered by ReAct agents, CodeAct frameworks, and advanced reasoning models.
AI-focused Computer Science Graduate
AI-focused Computer Science graduate specializing in AI Agent and Chatbot development and LLM-based Agentic systems. Hands-on experience building ReAct and CodeAct agents for competition-level code generation, automated reasoning, and multi-step task execution. Proficient in Python (LangChain, LangGraph, DSPy), RAG architectures, and agent orchestration frameworks. Demonstrated success in developing intelligent agents that leverage tool use, chain-of-thought reasoning, and action planning. Passionate about advancing autonomous agent capabilities and building production-ready AI systems that solve complex real-world problems.
Independent Research Projects
IEEE CSB Chapter, Green University of Bangladesh
IEEE Student Branch, Green University of Bangladesh
Office Daudkandi Pourashava (4, 5 & 6 Ward), Daudkandi, Cumilla
Built end-to-end LLM-powered pipeline for automated conference paper review and evaluation from PDF files. Implemented PDF text extraction, advanced prompt engineering for structured assessments, and intelligent reviewer assignment using LLM-based keyword extraction.
Developed autonomous coding agent using CodeAct architecture for competitive programming problem-solving. Implemented iterative code generation, execution, debugging, and refinement loops with LLM-powered reasoning. Achieved successful automated solution generation for algorithmic challenges.
Ranked top 10 for Bangla Language Processing Workshop 2025 with pass@1 score of 94% and 71%. Built bilingual (Bangla-English) LLM agent system for complex text understanding and generation tasks. Implemented agentic workflows with multi-step reasoning pipeline for improved accuracy.
Conversational chatbot powered by open-source LLM models with advanced RAG capabilities. Features context-aware conversation management, agentic reasoning workflows, and seamless integration with knowledge retrieval systems for enhanced response accuracy.
Laravel-based platform for interactive programming learning with role-based access control, quiz system, progress tracking, and multi-language course support with integrated Online Judge for code evaluation.
Responsive Laravel-based e-commerce platform with user authentication, shopping cart, order management, admin panel, secure routing, and dynamic content rendering. Live deployment showcasing full-stack capabilities.
CLI-based Python project using machine learning for detecting anomalies in datasets with rule-matching capabilities. Implements anomaly detection algorithms for network security analysis.
Java Swing desktop application using socket programming and multithreading for real-time client-server messaging. Supports multiple concurrent clients with efficient message broadcasting.
Green University of Bangladesh
CGPA: 3.50/4.00 (1 Semester remaining)
Researching RAG model optimizations for knowledge graphs, focusing on improved retrieval pipelines and embeddings to enhance temporal reasoning and answer accuracy.
Supervised by:
Professor Dr. Md. Saiful Azad
Dean, Faculty of Science and Engineering (FSE), Green University of Bangladesh
Accepted (November 6, 2025) — BLP Workshop @ AACL-IJCNLP 2025 (A* Conference)
Authors: Jahidul Islam, Md Ataullha, Saiful Azad
Paper explores iterative self-correction mechanisms and multilingual agent frameworks for improving Bangla-to-Python code generation tasks. Demonstrates novel approaches to enhance code generation accuracy in low-resource languages through agentic workflows and self-refinement strategies.
Python with Django
EDGE - ICT Division
July 2025
PHP with Laravel
BASIS-BITM
August 2023
1st Runner-Up - IT Olympiad
GUB CSE Carnival 2024
IEEE Membership
2024, 2025
Fluent (Reading, Writing, Speaking) - Native
Fluent (Reading, Writing, Speaking) – WSDA Certified
Dean, Faculty of Science and Engineering (FSE)
Professor, Department of Computer Science and Engineering
Green University of Bangladesh
saiful@cse.green.edu.bd
Lecturer, Department of CSE
Green University of Bangladesh
Ex-AI Engineer at Synesis IT PLC
+880 1879-093418
ataullha@cse.green.edu.bd
Research Engineer
Faculty of Science and Engineering
Green University of Bangladesh
+880 1746-555128
masud.re@green.edu.bd
Let's connect and discuss opportunities in AI agent development