Give every company the analysis a data science teamwould produce faster, cheaper, and without needing to hire one.
FOUNDER & BUILDER
Before data science, I served five years in the United States Marine Corps joining right after high school and was honorably discharged in 2019. The Marine Corps taught me what sustained, disciplined effort actually looks like over several years, not weeks, and it's the standard I hold this work to.
I immediately started college once I got out. Out of curiosity, I started learning Python & machine learning in early 2020. I started to really love the process of trying to solve difficult problems that didn't always have a clear answer. I spent the vast majority of my time in college working on personal projects mostly centered around tech and finance. This helped me secure an internship at JPMorgan Chase as an AI/ML analyst. At the end of my internship I received a return offer which I accepted. I graduated in 2023 earning my bachelor's degree in data science from the University of North Texas and joined JPMC as a data science analyst. I continued to spend my free time building projects in the evening and on weekends.
Eventually, after lots of testing, research, and experimenting I settled on the project you're viewing now. Seeing and experiencing the advancements of LLMs I started to notice a trend. Everyone was building copilots, AI that assist you in your workflow. This made sense with the limitations of LLMs. When I started building multi-agent systems, I wanted to give them the ability to improve themselves. With the vast amount of information and the speed of advancement in AI I knew that I couldn't possibly keep up with every new advancement. So I thought, what If the most advanced LLM models could help me help them. Over several months of building I started learning that It wasn't as simple as telling AI to improve itself, but with the right architecture and reporting structure, an LLM model could tell me what it needs to perform better. Whether that's a new data feed, structural improvements, or even additional AI agents. It took some time but I started seeing that an AI agent can not only see room for improvements but also realize what worked and what didn't. This type of memory recall started improving my multi-agent systems faster than I could on my own. I never allowed AI to actually implement these changes on it's own, I reviewed all suggestions and implemented improvements where I saw fit. This what eventually led me to Huie AI. The autopilot that provides the end to end analysis and the copilot that operates as the managing director telling me what I'm still missing. This is where I saw AI going, providing the end product instead of another AI assistant that helps you get there. So after two and a half years at JPMC, seeing what works and what doesn't, I decided I was going to go all in on this idea.
Sure, it's a very difficult endeavor but I'm drawn to problems that reward patience. The kind where the payoff comes from consistency, not shortcuts. Huie AI isn't something that will stay stagnant in it's current state. It's AI that will continuously grow and improve with time. I hope you follow my journey and enjoy what I build.
Connect on LinkedIn