Shafayetur Rahman

A structured AI and machine learning roadmap with topic-by-topic PDF notes for learners

A multi-agent research assistant built with LangGraph and MCP. Planner, researcher, and writer agents coordinate through a shared state graph — the researcher connects to a self-built MCP server for live web search via Tavily, and all LLM calls run on Groq (Llama 3.3 70B).

Clean NLP project for Quora Insincere Questions Classification with TF-IDF baseline, BiLSTM, GRU-Attention and Sentence Transformers models.

An end-to-end time-series forecasting pipeline utilizing Pandas optimization, feature engineering, and gradient boosting to predict 28 days of daily retail sales for the Kaggle M5 Walmart dataset.