T Theonic.Edition
Journey Artifacts Resume Let’s talk
THE PROGRAM
INSTITUTION
MIT Professional Education
DATES
Jul 7 – Aug 31, 2026 · 2026 cohort
SHAPE
8 modules · 1 executive capstone
THIS PAGE
Published module by module as I go — not assembled at the end
3/ 8 weeks
3 of 8 modules underway

Learning agentic AI
in the open.

Every module ends with something I can defend in a leadership room — a design, an agent, or a proposal. Those artifacts are the point; the certificate is a by-product.

LATEST · WEEK 3 · ASSIGNMENT 3.1
Conceiving & Programming of Agents
Designed an MCP-brokered launch governance agent for a frontier AI lab — architecture, guardrails, and a measurement plan.
READ THE WRITE-UP →
Read the latest — Launch Governance Agent → All artifacts See the eight-module outline
ARTIFACTS

Built, not just studied.

03 artifacts · one per module

Real things published as the program runs. Each one is a module’s assignment taken further than the rubric required. This page is itself one of them.

NEWESTChat → Agent → MCP → the launch stack
WEEK 3 · ASSIGNMENT 3.1
Launch Governance & Intelligence Agent
An agentic layer over the frontier-model launch process — filers, reviewers, and the approvals between them. →
ASSIGNMENT 2.1This page —
built by describing it.
WEEK 2 · CASE STUDY
Vibe coding the journey page
The page you’re reading is the artifact: generated as a mockup from a plain-language prompt, then prompted into working HTML. See how it was built.
COST PER 1M LOOKUPS
$20lightweight model
$1,500premium model, same answer
$0REST API — and 100% accurate
WEEK 1 · ASSIGNMENT 1.1
When the answer is not to use a model
Token-cost model for a lottery-results assistant at 10k and 1M queries a month. Premium reasoning cost 75× more for identical output — and a deterministic lookup shouldn't touch a model at all. Recommended a REST API with the LLM as an optional chat wrapper.

The eight-week
learning journey

/ Jul 7 — Aug 31, 2026

Each module builds toward the same frontier: deploying autonomous agents that create measurable business value — responsibly, and at enterprise scale.

Complete In progress Ahead
01 · Week 1Jul 7–13
✓ COMPLETE
Foundations of Generative & Agentic AI
What separates autonomous agents from generative models — and why this shift rivals the move to cloud.
The last platform shift was cloud. This one is agents.
02 · Week 2Jul 14–20
✓ COMPLETE
Models, Infrastructure & Trade-offs
What different models can and can’t do — cost, performance, and choosing the right tools for the enterprise.
Capability is cheap; the right fit is the hard part.
03 · Week 3Jul 21–27
IN PROGRESS
Conceiving & Programming of Agents
Designing an agent end to end — use case, integration path, guardrails, and the metrics that prove it worked.
Capability comes free; the boundary is the design work.
READ THE WRITE-UP →
04 · Week 4Jul 28–Aug 3
UP NEXT
Building Your First Agent
Hands-on prototyping of a task-scoped agent in a safe sandbox — from goal to tool use to evaluation.
Planned: turn the launch-governance concept into a working prototype.
05 · Week 5Aug 4–10
AHEAD
Integrating Agents into Ecosystems
Connecting agents to existing platforms and APIs so AI behaves as a cohesive layer across the org.
Planned: an MCP server map for a real legacy tool estate.
06 · Week 6Aug 11–17
AHEAD
Governance, Ethics & Responsible Scaling
Observability, model security, and the governance cadence needed to scale agents responsibly.
Planned: a responsible-scaling checklist for agent deployments.
07 · Week 7Aug 18–24
AHEAD
The Economics & Energy of Agentic AI
ROI modeling plus the rising cost and power footprint of running agents at enterprise scale.
Planned: a cost-per-decision model for the governance agent.
08 · Week 8Aug 25–31
CAPSTONE
Executive AI Adoption Plan
Synthesizing the program into a board-ready, agent-based transformation initiative.
The plan is the proof you can lead the change.
Capstone · Week 8 · Due Aug 31

Where all of this is heading

The program ends with a transformative, agent-based initiative for a specific business function — supported by an integration architecture, a detailed risk-benefit analysis, cost implications, and measurable KPIs that demonstrate strategic value.

My working thesis: the launch-governance agent scales into a governance operating layer — the same MCP-brokered pattern applied to every high-stakes approval chain in the enterprise. Week 8 is where I make that case in board language.

Integration architecture Risk-benefit analysis Cost model Measurable KPIs
REMAINING SCHEDULE
Building your first agentJul 28
Integrating into ecosystemsAug 4
Governance & responsible scalingAug 11
Economics & energyAug 18
Capstone submissionAug 31

Skills & tools in play

What the program is building on top of an existing transformation practice — strategic, technical, and governance.

Strategy
AI adoption roadmapping
ROI & KPI modeling
Change management
Executive communication
Technical
Agent orchestration
LLM selection & evals
MCP / API integration
Prompt pipelines
Governance
Responsible scaling
Model security & observability
Ethics & risk analysis
Data privacy
Notes from the weeks so far

“The last platform shift was cloud. This one is agents — and the risk is standing still while competitors move from efficiency to transformation.”

Reflection · Week 1

“Describing an interface well turned out to be the engineering skill. The model wrote the HTML; the specificity was mine.”

Reflection · Week 2

“Conceiving an agent is mostly deciding what it is forbidden to do — precisely enough that an engineer can build it and a lawyer can sign it.”

Reflection · Week 3
RESUME / CV

Hiring for AI transformation & governance?

Read the full résumé or connect on LinkedIn.

View full résumé → View LinkedIn