jthomas.site// notebook · v.4.2026
Consulting · AI/ML Advisory

AI & ML consulting for teams building at scale.

7+ years shipping AI infrastructure at Google, Amazon, and high-growth startups. From GenAI strategy to production LLMOps — I help engineering teams solve hard problems faster. Advisory, embedded, or project-based engagements.

Start a conversation Accepting select engagements
7+
years · Google & Amazon
10B+
events / day in production
1
US patent · swarm AI
27
academic citations
§ services 06 offerings

What I help with.

01 · Strategy

GenAI Architecture & Adoption

End-to-end guidance on adopting LLMs, RAG pipelines, and agentic workflows. Technical architecture reviews and build-vs-buy analysis for teams moving from prototype to production.

02 · Platform

LLMOps & ML Infrastructure

Production-grade ML platforms: prompt versioning, evaluation frameworks, model serving, and continuous fine-tuning pipelines. Hardened patterns drawn from Google's own internal stack.

03 · Data

Synthetic Data & RLHF

High-fidelity data generation pipelines for instruction tuning and alignment. Gemini-powered workflows for SFT and RLHF — the same techniques I work with daily on production training pipelines.

04 · Infra

Data Platform & Pipelines

Scalable data ingestion and real-time event streaming, data lake design, and warehouse architecture. Battle-tested patterns from building systems processing 10B+ events/day.

05 · Diligence

Technical Due Diligence

AI/ML technical assessments for investors and acquirers. Architecture reviews, risk analysis, team and codebase health, and competitive benchmarking with clear written deliverables.

06 · Speaking

Talks & Workshops

Keynotes, technical talks, and hands-on workshops on GenAI, swarm intelligence, and building AI at scale. Previous venues include IDSIA (Switzerland) and the University of New Brunswick.

§ engagement 03 models

Three ways to work together.

One-time

Advisory Call

90 minutes

Focused deep-dive into your architecture, roadmap, or a specific technical problem. Includes a follow-up written summary with actionable recommendations.

Monthly retainer

Embedded Advisor

Ongoing

Weekly sessions plus async access for architecture reviews, code reviews, and hands-on prototyping. Deep integration with your team's existing workflow and tools.

Fixed scope

Project Sprint

2 – 8 weeks

End-to-end delivery for a bounded problem — architecture, implementation, documentation, and knowledge transfer to your team. Fixed scope, fixed timeline, fixed price.

§ process 04 steps

How we work together.

  1. Discovery Call 30 min · free

    Quick introduction and overview of your current challenge or goal. High-level assessment and initial thoughts on approach. Determine if there's a mutual fit for a deeper engagement.

  2. Proposal & Scoping within 48 hrs

    Written proposal with objectives, deliverables, and milestones. Pricing options — fixed scope, hourly, or monthly retainer. Revision cycle until we both feel confident in the plan.

  3. Execute & Deliver duration varies

    Regular check-ins via your existing tools (Slack, GitHub, Jira, Linear). Iterative delivery with weekly demos and progress reports. Code reviews, pair programming, and architecture guidance as needed.

  4. Handoff & Support +2 weeks

    Full documentation, runbooks, and architecture decision records. Team walkthroughs and Q&A sessions on the delivered work. Two weeks of post-engagement support included at no additional cost.

“Architecture compounds. Make the load-bearing calls early.”
working with teams shipping to production · limited availability