I'm Seamus, and I hold a Master's in AI, where I build things instead of just studying them. My work sits at the intersection of two things: shipping real AI products, and understanding what actually makes them work under the hood. On the build side, I've worked on AI agents using LangGraph, Groq, and Pinecone which includes a scheduling agent that turns plain-English requests into full calendar plans, not just simple task entries. On the research side, my dissertation compared LLM alignment methods (DPO, ORPO, KTO) that tests how different fine-tuning approaches actually perform, not just what the papers claim. I've also published research on sentiment analysis for Konkani, a regional Indian language. Right now, I split my time between building and helping other people cut through the noise around AI. There's a lot of "just use this framework" advice out there that doesn't account for what a specific problem actually needs like does it need a no-code tool, a custom agent, or sometimes just a smaller fix than people expect. If you're trying to automate a workflow, build an AI agent, or just figure out where to start, book a call. I'll give you a straight, honest read on your situation this includes telling you if you don't need what you think you need.