I build AI products and help teams put AI to work.

Now

At DealMaker, I help teams across the business put AI to work. I love how much of the organization I get to work with: understanding what people need, building useful tools, and helping them use those tools well.

I joined as a Senior AI Engineer in September 2025 and became Director of AI & Infrastructure in March 2026. I still build in Ruby and Python, with a focus on agents, retrieval, and the evaluations that tell us whether a system is doing its job.

I’m currently exploring inference engineering and clearer communication. Away from work, I’m cooking, playing Magic: The Gathering, and tinkering with a PSP.

What I’m building

My story

I came to AI through software consulting and a habit of following questions. Consulting taught me to understand the problem before reaching for a tool. AI gave me a whole new set of questions to chase.

Read the full story →

Landon Notes

Building AI products. Helping people use them. Sharing what I learn.

Practical lessons on engineering, adoption, and the decisions in between. Read the latest notes below, or get substantial articles and occasional roundups by email.

• April 30, 2026Alan Weiss and Knowing Your Worth
• March 15, 2026A Non-Engineer's Guide to Building Autonomous AI Agents with Claude Cowork
• March 1, 2026Dario Took a Stand

Browse all notes → · Get Landon Notes by email →

Selected production work

• Making an investor-verification agent observable and testableDealMaker · 2025–26Migrated the 506(c) AML verification agent from n8n to Google ADK, comparing behavior in shadow mode before cutover. Built the evaluation workflow with domain experts reviewing traces and annotating failures, and made every run searchable and replayable in Datadog. Held end-to-end latency under ten seconds.
• DealMaker Labs: an internal AI tools programDealMaker · 2026Designed the operating model end to end: employee ideas move through build support to managed production hosting, backed by AI governance guidelines and cost tracking. Seven employee-built tools were registered across Sales, Marketing, Operations, and Broker Dealer in the first week. Read the launch post →
• AI search for vehicle and RV dealershipsAutomotive · 2024–25Built a RAG-based search system that lets shoppers query inventory conversationally. Replaced subjective quality checks with automated RAGAS evaluation as a release gate, and profiled the pipeline to cut response times.

Training & projects

• AI-powered search trainingHealthcare · 2024 · open-sourcedTrained 14 healthcare engineers in NLP, vector search, LLMs, and RAG, with healthcare’s regulatory considerations built into the material. I then open-sourced the course so any team can use it. Course materials · Labs

  

• Forecasting the FutureMachine learning tutorials assumed Python. I built a weather-prediction model in pure Ruby and presented it at RailsConf 2023. It shows ML fundamentals don’t depend on the language. Talk · Code
• SimpleRAGRAG tutorials started with a heavy stack. I built the whole architecture in one small gem: no vector database, a few light dependencies, an API key. It makes each stage of retrieval readable. Code
• Ruby ReAct AgentAgent frameworks hide the loop you most need to understand. I implemented ReAct directly: reason, act, observe, with real tool access. It shows how an agent works before you trust one in production. Code

More on GitHub.

Founder of the RubyAI community

I founded and run RubyAI, a community on X for engineers building AI applications in Ruby.

Everyone was building AI in Python. I set out to prove what Ruby could do, and to raise AI literacy across a community I care about.

Speaker

My first talk was Forecasting the Future: Machine Learning for Weather in Native Ruby at RailsConf 2023 in Atlanta. I pitched one talk to two conferences and got into both, then spent weeks of evenings turning a weather-prediction model into pure Ruby. The room filled and people were turned away at the door.

Since then (see more on RubyEvents):

• —Ruby on RAG: Building AI Use Cases for Fun and Profit, Madison Ruby 2025
• —An Intro to Machine Learning with Ruby, Blue Ridge Ruby 2023, Asheville
• —Building an AI Community with Ruby, Flower City AI 2023, Rochester
• —The Missing Piece: Tooling for Robust AI in Ruby, Ruby Warsaw 2023
• —Surely You Must Be Joking, Jupyter Notebooks with Ruby!, Boulder Ruby 2023, which was shared on Hacker News

I helped plan RubyConf 2023 in San Diego as a member of the Ruby Central program committee, selecting conference track topics and reviewing Call for Papers submissions.

Podcasts

I’ve been a guest on:

• —What happens when the model can’t fix it?, freeCodeCamp #213 with Quincy Larson, 2026
• —An aberrant generation of programmers, Changelog & Friends, 2024
• —AI Consulting with Landon Gray, Indie Rails #41, 2024

Work with me

My independent focus is Shinkiro. I’m also open to selected advisory engagements on AI strategy, product direction, and adoption.

Have a decision you’re working through? Email me with a little context.

Contact

For advisory work, talks, podcasts, or training, tell me what you’re building and where it’s stuck.

Email landonnospam.gray (at) hey (dot) com, or find me on LinkedIn, GitHub, X, and Bluesky.