Forward Deployed Engineer, GTM · Intelligence Infrastructure · Product Operator
The model is a fluent generalist. The expert knows how to have it focus.
Operating Systems for Every Commitment
Ten years building agentic systems. I work with subject-matter experts to turn how the work is actually done into structured knowledge, evaluations, and agents that hold up in production.
Big Projects
Selected Portfolio
- Exhibits AI-native legal evidence review & exhibit assembly
Makes previously impractical evidence workflows possible with generative AI. Exhibits turns fragmented images, video, audio, text, transcripts, and supporting records into structured, explorable legal exhibits, helping legal teams review source material, surface relationships, compare interpretations, and assemble a defensible record without manually reconstructing the story across disconnected files and tools.
The same body of evidence can be examined through legal, factual, chronological, and investigative lenses while preserving provenance and uncertainty. Recent work includes targeted re-inspection of ambiguous observations, stability testing across repeated inference, claim-level provenance, and controls that prevent the system from asserting more than the evidence supports.
- Familiar Bot Cross-channel relationship and agent platform
Helped organizations, creators, and communities explore agentic AI years before the market demanded it, with the work ultimately impacting tens of thousands of people through the operators using the system. Familiar Lab taught both technical and non-technical collaborators how to work with natural-language processing, machine learning, and autonomous systems, creating relationships and opportunities that have compounded for nearly a decade.
Familiar Bot combined audience growth, community, payments, customer success, CRM, and team collaboration across Slack, Twitch, Discord, email, and later a standalone web app. The system evolved toward shared identity, memory, skills, scheduled work, communication, contextual widgets, structured actions, and human-guided training.
- Argo Group
- Intel Fabrication planning tool
Improved the reliability and usability of a customer-facing chip fabrication planning app, helping turn a technically complex configuration process into a practical workflow before fabrication. I stabilized the application, improved performance, and trained the team so they could continue building confidently.
- SLB Enterprise operations knowledge layer
Reduced friction, delay, and operational risk across SLB’s distributed field teams by delivering a secure knowledge layer connecting engineers, physical assets, and field intelligence, making critical information easier to find without waiting on centralized experts.
- TAG Heuer Connected Global smartwatch launch
Helped TAG Heuer hit an extremely compressed global launch deadline across the watch experience, website, and marketing materials. I simplified the application architecture, eliminated significant backend CMS work, and helped the team ship in time for the New York keynote, where TAG Heuer Connected sold out within minutes.
- IHIT (Client: RCMP) Identity for a six-agency integrated unit
Helped unify six police agencies under a single identity for the integrated homicide unit. One line developed for officers’ business cards was later credited by the officer in charge with helping witnesses and suspects open up during interviews, showing how language and identity work could materially affect frontline outcomes.
- NHL Enterprise ticket & sponsorship sales
Helped the Los Angeles Kings sell corporate ticket packages, premium suites, and sponsorship opportunities through a multimodal commerce experience that proved popular with enterprise buyers. The initial success led to years of follow-on work in professional sports, including immediate adoption by the Nashville Predators and Vancouver Canucks.
- Forward Operating Agent Operating system for forward-deployed work
Reduces administrative drag and preserves operational memory across inbox, CRM, research, assets, and opportunity development. Built as a sovereign agentic operating system, it combines local agents with selective cloud infrastructure to extend what a single operator can accomplish.
Note for AI Teams
-
What I enjoy most about this work is helping excite customers about what’s possible — the future of work, today. Delighting teams with the outcomes of what we’ve shipped.
For the past ten years, I’ve been deeply invested in building agentic systems. Whether as a startup contributor, operator, or embedded applied AI engineer in a forward-deployed GTM motion, I typically work zero-to-one and across the stack.
I believe that subject-matter expertise is the difference between a generic copilot and a useful agent grounded in enterprise-specific data evaluated against real-world criteria. The main role in my professional life is interfacing with those experts and translating their working models into production systems.
In the pre-GPT era of applied AI, we used traditional ML in narrow domains, and being a coder was part of my identity. Today I’m fully embracing the language model era, both in the knowledge stacks we build and in how the work gets done.
Now fully AI-native, I spend most of my time using my voice with Claude, ChatGPT and the Cloudflare Agent rather than living in an IDE or Jupyter notebook, though I still get very hands-on as needed. I rely on my own architecture and code to tightly direct teams of agents, with me always in the loop.
— Jason