Farhad Rezazadeh

A multi-agent reinforcement learning framework for a custom OpenAI Gym environment (BS-v0), combining Echo State Networks (ESN) for temporal encoding and Graph Convolutional Networks (GCN) for relational state representation, trained via the REINFORCE algorithm.

Harnessing Retrieval-Augmented Generation (RAG) with LangChain for intelligent and efficient AI-powered information retrieval and decision-making.

This application allows users to engage with multiple AI models (currently including OPEN AI, CLAUDE, and OLLAMA) simultaneously.

Pytorch-Tutorial is a structured, beginner-to-advanced series of Jupyter notebooks that walks you step-by-step through the core concepts, workflows, and best practices for deep learning with PyTorch.

A complete, end-to-end educational course on building scalable LLM and Agentic AI systems using Modal’s serverless compute platform with source code, notebooks, and examples for full YouTube course.

Masked‑patch Transformers for multivariate time‑series: self‑supervised pre‑training, KPI fine‑tuning and lightning‑fast inference — all in pure PyTorch.