About

People ask me what I actually do all day. My answer: a little bit of everything.

In my day to day, I am an Area Technical Lead at Google Cloud, overseeing developer relations, information experience, and frameworks and languages while focusing on how applied generative AI changes the software development lifecycle. Before this role, I led Developer Advocacy for Data and Analytics at Google Cloud, working with engineers in Python, SQL, BigQuery, and Spanner. When I speak at conferences (Google Cloud Next, PyCon, or Ai4 as example), I will usually vibe code a custom interface live rather than show production screenshots.

How I got here

Fun fact: I did not start in computer science. Not even close.

In one of my first roles out of college, I was a process analyst and change management consultant, value stream mapping client workflows to find where the drudgery lived. I designed close to 100 trainings before I ever touched production software development. That detour stuck with me. It is why I still care so much about how teams transfer the 'why' behind a system instead of just copying commands from readmes / wiki's.

From there, I became a data engineer at heart. Before joining Google, I served as a Practice Director at a regional consultancy, leading software, data, and DevOps teams through messy migrations where org charts swing like pendulums and someone still carries the pager.

What I build on nights and weekends

Let's be realistic. Slide decks do not ship code.

To me, being 'all in' on AI means defaulting to models and agents as your first partner, then fixing the machine when they trip. I run 100+ agent sessions a week. I rely on my developer experience to build mini systems of improvement so each project gets better over time (usually). When a question bugs me, I build a working prototype under my What I Built This Month series:

• cosm and topocosm: an agent-first source control system using AST's and Merkle trees, built because multiple agents writing to one repo quickly break git and burn tokens scanning thousands of lines of code.
• Electric Sheep: a long running agent memory system backed by Google ADK, Gemini, BigQuery, and Cloud Spanner so agents consolidate 'dreams' across sessions.
• smartrouter: a local and cloud worker router. I built it to build good things for an economic cost.
• session-journal: a sidecar skill that summarizes my chat logs in .local, tracks where agents failed, and keeps me from losing essay ideas.
• Warehouse Dash and Aetheria: a 10-tick-per-second multiplayer forklift game on Cloud Run and a Three.js sandbox.

Life outside the terminal

The reality: I also like building things that do not compile.

I live in Colorado. The home renovation part of my tagline is very real. Away from the terminal, I am usually tearing into a house project or working on a rental property near our place. Recently, I had to fix a massive, shadeless front yard during a brutal Colorado drought without paying a landscape architect thousands of dollars. I fed a photo of the yard to Gemini with my low water requirements, got back a xeriscape plan, and did the shovel work myself. AI did a 'C' job. For yard work, that translates to 'good enough'.

Ultimately, whether it is cloud architecture, agent harnesses, or hauling rock in the Colorado sun, I just like figuring out how the machine works and removing the toil. For companies and engineering leaders, if you want to talk about developer workflows, agentic systems, or conference speaking, reach out on LinkedIn. For the individual developer, poke around the code, try the forklift game, and hit me up in the comments!