Washington, DC

2026 / 05 / 24

Daniel Fenton.

AI platform leader and builder. I build governed agent platforms at enterprise scale and still write code.

Currently

Leading enterprise AI adoption · shipping agents and MCPs · cooking with Chef Claude
§ 01 · Thesis & summary
DWG / INTRO
“In 2023, GPT-4 came out and I realized it could code. AI became the first technology that ever scared me, so I learned to build with it. Businesses face the same choice: master it, or get left behind.”

NYU Stern · Mar 2026 · “The Wave Is Coming”

I build JLL's enterprise AI platform for a 110,000+ employee global workforce. I work across product strategy, platform architecture, AI governance, adoption, and code. I build directly with Claude Code and Codex — and teach the people around me to do the same.

Two industries — commercial real estate and civic tech — with platforms as the consistent unit of work. Before JLL-GPT I launched ArcGIS Hub at Esri (now 27,000+ sites). Before that, Blackbird at JLL — one of the first analytics platforms in commercial real estate, still in service a decade later.

The rest of this site is the work that doesn't fit on a resume.

§ 02
ASSY / NOW

What I'm up to

At work
Day job · the big platform

Day job · 2023 — present

JLL AI Platform

I lead JLL's enterprise AI platform strategy and product execution: JLL-GPT and Falcon, the platform underneath. It serves a 110,000+ employee global workforce. I work across product strategy, platform architecture, AI governance, adoption, and code.

The architecture: tiered agents, an MCP gateway for tools, an LLM gateway routing across AWS, GCP, and Azure models, RAG, governed Databricks and Unity Catalog access, auditability, policy, and safe deployment paths.

Solo builds alongside the leadership work: a governed document-review agent for a sensitive enterprise workflow, moving from prototype to production in six weeks; Pulse; and the JLL work-tools MCP that connects the AI platform to work context across everyday tools.

JLL Falcon announcement →

Fig. 1 — Platform topology Sheet 1 / 1
JLL-GPT Pulse Assistants Work tools MCP GATEWAY · LLM GATEWAY MODELS DATA AWS GCP AZURE frontier LLMs MS365 DATABRICKS enterprise data AZURE · K8S · UNITY CATALOG · GUARDRAILS RUNTIME · GOVERNANCE · SAFETY APPS CORE INPUTS BASE
Apps over a gateway. Models routed across clouds, data pulled from the enterprise, all governed.
Side projects
Two agents I'm running at home

On nights and weekends I ship Claude agents for my own life: a kitchen, a drawing studio. Each is a long-running agent with a phone in its pocket and a job to do.

9:41
● ●●●
Chef

Week of May 24

7 MEALS · 1 GROCERY RUN
MON Sheet-pan harissa chicken ·
TUE Miso-glazed cod, rice ·
WED Black bean tacos ★
THU Cacio e pepe ·
FRI Salmon, asparagus ·
SAT Margherita pizza night ★
SUN Roast chicken, schmaltz potatoes ·
GROCERY LIST 18 items →
FIG. 2 Chef Claude — weekly meal plan

Chef Claude

2026 —

A personal chef for the Fenton household. A long-running Claude agent picks the week's recipes and the grocery list. It takes feedback and learns what we actually like. We cook from it almost every night. Native iOS on the phone, FastAPI on AWS underneath.

~5/wk meals planned 2pax users

github.com/dmfenton/chef →

9:41
● ●●●
Monet

Plot #034

ITERATION 12 / 16 · ON PLATEN
Untitled study, generated by Code Monet
AGENT LOG
19:42 perceive captured iter 11
19:43 replan widen field, +12%
19:43 actuate queued strokes 73–96
19:48 perceive looking…
REPLOT
APPROVE
FIG. 3 Code Monet — agent in mid-iteration

Code Monet

2026

Open-source autonomous AI art system. Claude agents write drawing code, inspect screenshots of their own output, and revise the work with persistent memory. A closed-loop agent: plan, draw, perceive, replan. Built solo, end to end — AWS, Docker, server, mobile, agent loop, multi-user with magic-link sign-in — all open source.

PLAN → ACTUATE → PERCEIVE → REPLAN

monet.dmfenton.net →source →

§ 03
BOM / IN-SERVICE

Things still running

Software I made or led that's still in production. The work I'm proudest of, mostly because it lasted past my interest in it.

  • Blackbird

    2010 — 2013 · 2019 — 2023

    Fig. — SF competitive landscape JLL Blackbird
    Blackbird's 3D San Francisco view: a focal property in red, competing assets in tan, within a configurable radius
    Focal asset in red, competing assets in tan, all inside a configurable radius. Watch the JLL video →

    Created it as a product manager at JLL. One of the first analytics platforms in commercial real estate, before "data platform" was a default category. Now in use globally at JLL. Came back later as Senior Director of Product Engineering to set its next chapter; the simplified 3D visualization we shipped delivered an immediate 50% increase in usage.

    pendo.io case study →

  • ArcGIS Hub & Open Data

    2014 — 2019

    Fig. — Hub deployments worldwide ArcGIS Hub
    27,843
    Hub sites
    82,026
    Apps & maps
    143,243
    Datasets
    Every dot is a live ArcGIS Hub site. Counts and locations fetched at build time from Esri's public search API.

    Launched ArcGIS Hub and the Open Data product line at Esri. It powers around 27,000 government sites today: counties, cities, agencies, all of them making public data browsable and downloadable. The District uses one (opendata.dc.gov, pictured); so do most of the others. I designed the search, picked Elasticsearch, built the Node.js ingest backed by Postgres, and ran the team that scaled global ingest from 18,000 downloads a month to over a million.

    opendata.dc.gov →

  • Koop

    2015 — present

    A developer-friendly API interoperability library: an open-source toolkit that turns any data source into a geospatial API. Shipped at Esri; still in use today.

    koopjs.github.io →

§ 04
NOTES / METHOD

How I work

Builder PM. I lead, and I build. Claude Code and Codex are the daily drivers.

Platform before feature: a central orchestrator, an MCP gateway, an LLM gateway, a tiered development model, then a governance fabric. Features come after the chassis works. Scenarios first. Every agent track is anchored to a concrete business scenario before code is committed.

A bias toward closed loops. The most interesting agents observe what they made, learn from it, and take the next useful step: a drawing revised on screen, a valve opening in the garden, a meal plan changing what we cook. Side projects pick a real feedback loop on purpose.

Communication: clinical, concise, substance over fluff. Slides and memos are precise strings, not paragraphs.

§ 05
MISC / WORKSHOP

A whole host of other fun stuff

Workshop · the rest of it

Roughly ninety more experiments live on GitHub.

github.com/dmfenton →
Agents

Long-running Claude agents for the kitchen, the garden, my reading list. Each one tries to close a real feedback loop.

Hardware

A GPS cat tracker on Helium, a wildlife camera, dev-board work on the AMB82-Mini NPU, HDR pipelines on Picamera2, occasional servo repair on a small robot.

Studio

Generative drawing, a Riso & sticker phase, and a healthy archive of experiments that didn't quite ship.

The building habit doesn't stop at software.

§ 06
LOG / TALKS

Talks

Industry keynotes & panels

Show more talks →

More industry keynotes & panels

  • 2026

    Trepp Connect NYC

    Panelist

  • 2024

    ULI Spring Meeting NYC

    Panel: How AI Will Transform the Office Experience in the Next Decade

Recurring

  • 2023 — present

    NYU Schack Institute of Real Estate

    Guest lecturer, multiple semesters

Builder & technical (archive)

Podcasts

§ 07
NOTES / WRITING

Writing

All writing →

§ 08
HIST / CAREER

Where I've worked

  • 2023 — present

    JLL · Head of Product, AI Platform

    Lead JLL-GPT and Falcon, the enterprise AI platform underneath.

  • 2019 — 2023

    JLL · Director, then Sr. Director of Product, Market Intelligence

    Led Market Intelligence product; returned to lead Blackbird's next chapter.

  • 2017 — 2019

    Esri · Lead Software Engineer

    Led search and data discovery for ArcGIS Hub.

  • 2015 — 2017

    Esri · Software Engineer

    Scaled ArcGIS Open Data ingest and downloads; created Koop.

  • 2014 — 2015

    Esri · Lead Product Engineer

    Led product design for ArcGIS Open Data from launch.

  • 2011 — 2014

    JLL · Research Analyst, Research Manager, Product Manager

    Built Blackbird, geospatial intelligence for commercial real estate.

Education

  • University of Maryland, M.S. coursework, Computer Science: geospatial computing, information retrieval, human-computer interaction.
  • Georgia Tech, MS, Applied Economics.
  • College of Charleston, BA, Economics and Philosophy.

Patent

US 9,892,474, Computing system and method for visualizing integrated real estate data. Issued 2018.

§ 09
META / CONTACT

Contact