MatthewF.Wilson,Ph.D.
Director of Research Translation and AI Strategy at Baylor's Institute for Global Human Flourishing, where I develop AI-powered instruments for wellbeing research.
What I Build
I work on AI systems that solve real problems. Most of my projects fall into a few categories.
GraphRAG Systems and Knowledge Bases
Most of my work involves building systems that let people query their own information. Maybe you have a large document library, a research corpus, internal wikis, or years of accumulated knowledge that's hard to search effectively. A well-built GraphRAG system makes all of that accessible through natural language questions, including the connect-the-dots questions that plain vector search tends to miss.
I pay a lot of attention to retrieval quality, because that's where most RAG systems fail. It doesn't matter how good your language model is if you're feeding it the wrong context. GraphRAG helps by extracting entities and relationships from your documents and pairing a knowledge graph with the vector store, so the system can traverse connections instead of only matching on similarity. I think carefully about how the graph is built, how chunking and embedding choices play against it, and how the retrieval combines graph and vector signals. Get any of those wrong and you get confident-sounding nonsense.
Workflow Automation
The other big category is connecting tools into automated workflows. Most businesses use a dozen different software products that don't talk to each other very well. I build the bridges.
I work primarily in N8N, Make, Supabase, and GoHighLevel, though I'll use whatever makes sense for the problem. The goal is always to create systems that run reliably without constant attention, not Rube Goldberg machines that break the moment something unexpected happens.
Prototypes and MVPs
Sometimes you have an idea and you want to see if it works before investing heavily. I enjoy these projects because they're creative and move fast.
I work agentically with AI tools, building iteratively rather than writing detailed specs in advance. For the right kind of project this gets you to a working prototype in days rather than weeks. It's not the right approach for everything, but when it fits, it's remarkably efficient.
Voice AI and Conversational Systems
This is an area I find genuinely interesting. Voice interfaces are getting good enough to be useful, and there are a lot of applications that make sense: phone systems, scheduling agents, conversational assistants for specific domains.
I've built systems that handle appointment scheduling, answer questions from knowledge bases, and guide users through multi-step processes. The technology is evolving quickly.
Current Project
Hello! I'm here to help you assess your wellbeing. Let's start with a simple question: How would you rate your overall life satisfaction on a scale from 0 to 10?
I'd say around a 7.
That's great to hear you're at a 7! Can you tell me more about what's going well in your life right now?
I've been spending more time with family and focusing on my health.
FlourishingSurvey.com
The project I'm most excited about right now is an AI-powered survey instrument for human flourishing research. Traditional surveys ask the same questions of everyone and move on.
This system uses conversational AI to ask intelligent follow-up questions based on how someone responds, which lets researchers go much deeper than a static questionnaire allows.
Visit FlourishingSurvey.comBackground
I've had a somewhat unusual path to building AI tools. I spent about a decade in corporate leadership roles at companies like Volvo and Danaher, then pivoted into academic research at Harvard's Human Flourishing Program and now Baylor.
The academic background means I think carefully about what I'm building and why. The corporate background means I care about shipping things that actually work. Both feel like advantages in this space.
Read More About Me