Abhijoy Sarkar is an AI engineer and researcher who builds agentic systems and large language models for a living. He is the co-founder of PromptGuard, an AI-security startup building a multi-agent safety engine for enterprise LLMs, and has advised teams at PwC and Fortune 500 companies on agentic automation and LLM deployment. At Multiverse Computing he built a thirty-billion-parameter enterprise LLM from scratch and co-authored published research on quantum-inspired model compression. Earlier he built voice-search AI at Flipkart used by hundreds of millions of daily users, and frontier LLM prototypes at TheoremOne. He holds a Master's in Artificial Intelligence from IIT Ropar. This course is the one he wishes existed when he started: agents explained by building them, from the loop outward, not by importing a framework. You will write every line yourself and leave able to read and extend any agent framework on the market.