Brett Hardiman

I started out at Johnson & Johnson Insurance as a Business Analyst — gathering requirements, writing specs, updating SQL databases, and learning how to get developers and stakeholders on the same page. That role gave me the foundation I still build on: how to listen to what people actually need, how to translate that into something a dev team can run with, and how software gets built when real constraints are involved.

I moved to Booz Allen Hamilton as a Requirements Analyst and ended up going deep fast — became the go-to person on several government applications, got my CSM and CSPO certifications, and started wondering what else I could do. I could see AI was about to change everything, and I knew that if I wanted to stay relevant — especially coming from a non-technical role — I needed to get more technical. So I enrolled in Booz Allen’s Data Science Tech Excellence Program and started learning Python, machine learning, NLP, and large language models from scratch. That bet paid off: I got promoted, moved to the AI/ML ART, led three AI initiatives from concept to MVP, and was hands-on reviewing data science outputs and contributing to architecture discussions along the way. Along the way I started building AI agents. My Tech Lead had architected a multi-agent pipeline for the team, and I built the Requirements Agent that plugged into it — my agent took his architecture output, broke it into development stories, and pushed them into the backlog for his downstream agents to pick up and work. Owning that node taught me how agentic systems actually fit together.

Outside the pipeline, I built three standalone agents using Jira MCP: a DSU facilitator, a sprint summarizer, and a Jira research agent that surfaced quick reports on initiatives — what they were, when they happened, who the stakeholders were, and who to contact. My team used them, I trained my colleagues on how to, and those tools are now part of how the team operates day-to-day.

I also wanted to prove I could own the whole thing end-to-end. So I built my own 8-agent SDLC pipeline from scratch — architect, requirements, coding, code review, security, CI/CD, project manager, and summary agents — used it to build this portfolio site, then extracted it into the Agent Pipeline Kit as a standalone reusable framework anyone can drop into a project. That’s the one I’m most proud of.

I think that is just truly how my brain is wired. I get curious about something, I go learn it, and I build something with it. Every project on this site started the same way — “what if I could…” — and turned into something that actually works. And through all of it, I’ve realized that making workflows more efficient, easier, and quicker isn’t just something I’m good at. It’s what I actually care about.