Agentic Engineering Field Curriculum

Published path · practitioner · approximately 9 hours

Engineering with agents without losing engineering

Move from model-assisted coding to a bounded, inspectable, independently challenged delivery system.

Designed for Engineer, Architect, Platform. 7 finite stages.

1. 01 · See the whole system
2. 02 · Align before implementation
3. 03 · Shape an agent-readable codebase
4. 04 · Slice for fast feedback
5. 05 · Build against evidence
6. 06 · Verify independently
7. 07 · Scale without losing control

Markdown edition

Published path · advanced · approximately 9 hours

Private AI and owned inference

Choose deliberately among hosted, dedicated, local, self-hosted, adapted, and trained model paths without treating deployment location as a complete privacy claim.

Designed for Engineer, Architect, Platform, Risk. 6 finite stages.

1. 01 · Trace where the data comes to rest
2. 02 · Choose how much of the stack to own
3. 03 · Run inference on hardware you control
4. 04 · Serve it like production infrastructure
5. 05 · Adapt only with evidence
6. 06 · Govern the model lifecycle

Markdown edition

Published path · advanced · approximately 10 hours

Agent platform engineering

Specify the identity, substrate, context, tool, orchestration, verification, evidence, and lifecycle services that bounded agents require, as contracts an independent reviewer can challenge.

Designed for Architect, Platform, Risk, Operations. 7 finite stages.

1. 01 · Specify the execution substrate
2. 02 · Give every agent standing
3. 03 · Version context and memory
4. 04 · Draw the tool boundary
5. 05 · Design durable runs
6. 06 · Design the verifier plane
7. 07 · Close the evidence and lifecycle loop

Markdown edition

Published path · practitioner · approximately 9 hours

Lead AI-native delivery

Redesign the operating system around higher generation speed while protecting strategic judgment, independent assurance, and outcome ownership.

Designed for Executive, Engineer, Risk, Operations. 6 finite stages.

1. 01 · Measure the system, not the activity
2. 02 · Find the absorption constraint
3. 03 · Redesign work and attention
4. 04 · Scale verification and evidence with generation
5. 05 · Bind governance to execution
6. 06 · Close the loop from production

Markdown edition

Independent practitioner source

Matt Pocock

Developer educator and creator of AI Hero and Total TypeScript.

Attributed positions

• Software fundamentals and strategic engineering judgment become more important as code generation accelerates.
• Teams should reach shared understanding before turning an idea into a specification and implementation plan.
• Vertical slices, tight test feedback, deliberate module boundaries, and fresh phase contexts make agent-assisted delivery more inspectable.
• Large, uncertain efforts benefit from resolving decision questions before they become implementation tickets.

AISDLC extensions

• A fresh-context review is a correlated quality check; independent verification requires separately owned challenge with an appropriate model lineage or deterministic method.
• Specifications and skills guide work but do not confer authority; governance assigns standing and the control plane makes constraints binding.
• Consequential release requires attributable evidence and a named human decision authority beyond an implementation workflow.

Sources and further study

1. “Software Fundamentals Matter More Than Ever” — Matt Pocock

   AI Engineer · A practitioner argument for retaining strategic engineering judgment, inspectable code, test feedback, deliberate module design, and human comprehension as generation accelerates.

   Practitioner perspective · speaker · link-or-standard-embed-only

   Limitation: This is an attributed conference position, not a controlled study, consensus standard, universal productivity law, or enterprise release method.

2. mattpocock/skills: A complete AI Coding workflow, end-to-end

   Matt Pocock · A walkthrough connecting idea interrogation, durable context, specifications, tracer-slice tickets, implementation, test-driven feedback, and code review.

   Practitioner perspective · creator · link-or-standard-embed-only

   Limitation: This demonstrates one practitioner workflow and tool ecosystem; it does not establish independent verification, governed authority, or fitness for every team and codebase.

3. /wayfinder: Nothing is too big to plan anymore

   Matt Pocock · A practitioner model for turning large, foggy efforts into explicit decision tickets whose resolved map can later become a specification and implementation plan.

   Practitioner perspective · creator · link-or-standard-embed-only

   Limitation: Wayfinder is an evolving practitioner method. Its planning map does not replace accountable product decisions, architecture review, implementation controls, or release evidence.

4. Software Design playlist

   Software Design · Harry Phung · A third-party public collection by Harry Phung / Software Design that includes the supplied Matt Pocock talk and other software-design material for editorial discovery.

   Practitioner perspective · featured-in-third-party-collection · link-or-standard-embed-only

   Limitation: Matt Pocock is featured in the collection but is not represented as its curator or publisher. A playlist is a discovery surface, not evidence for a claim; individual items require review and attribution.

5. Skills for Real Engineers

   GitHub · mattpocock/skills · A public, evolving repository of agent skills and workflow documentation for engineering, planning, testing, review, teaching, and maintenance.

   Open-source implementation reference · author-maintainer · mit-source-with-attribution

   Limitation: The repository is a practitioner implementation, not an enterprise control standard. Substantial reuse must preserve its MIT copyright and permission notice.

6. AI Hero

   AI Hero · Matt Pocock’s canonical public teaching site for AI engineering articles, videos, workshops, skills, and cohort information.

   Practitioner home · canonical-home · link-or-standard-embed-only

   Limitation: Public descriptions can establish attribution and available offerings, but paid lessons, course exercises, visual assets, and brand identity are not AISDLC content.

Relationship boundary: Matt Pocock is included as an independent practitioner source. AISDLC owns this synthesis and does not represent him as a contributor, partner, endorser, or approver.