The Codified Practitioner

AI didn't make engineering judgment obsolete. It made it the bottleneck.

A program for senior engineers who want to convert twenty years of accumulated judgment into AI systems that scale their work. Built around a methodology, not a prompt library.

Apply for the founding cohort →Read the methodology

The Problem

The senior engineer's AI moment is here, and most of the advice is wrong.

Something has shifted in the last twenty-four months. Performance reviews are starting to mention AI fluency. Internal initiatives are forming and senior engineers are noticing they weren't invited. Twenty-eight-year-olds on the team are shipping faster with Cursor. The CTO did an AI strategy all-hands and the senior people couldn't have contributed to it intelligently.

If any of this sounds familiar, the standard advice is to "learn AI tools." Take a course on Copilot. Read about prompting. Pick up a new stack.

That advice misreads the situation. The senior engineers who feel behind aren't behind on tools — they're sitting on twenty years of accumulated judgment, debugging instincts, architectural taste, and domain expertise that almost nobody else has. The tools aren't the problem. The problem is that this judgment lives in their head, where AI can't reach it.

Until the expertise comes out, AI just makes everyone faster at the wrong things.

The Reframe

Your expertise is the raw material. AI is the multiplier.

The senior engineers who are winning the AI shift aren't the ones who learned the tools fastest. They're the ones who figured out how to take what's in their head and put it into systems.

When a senior engineer evaluates a pull request, they're running an internal checklist they couldn't fully articulate if asked. When they debug a production incident, they're pattern-matching against years of previous incidents that left no written trail. When they make an architectural call, they're weighing tradeoffs they've internalized to the point of invisibility. That's the expertise.

The work of codifying it — making it explicit, structured, and operational — is engineering. It has stages, deliverables, and patterns. It is learnable. Most senior engineers haven't done it because nobody has shown them how. Once they do, their work changes.

AI doesn't replace the engineer who can do this. It amplifies them by ten or a hundred times. The judgment was always the bottleneck. AI just made it visible.

The Methodology

Eight stages from tacit knowledge to operational AI systems.

The Codified Practitioner methodology walks you through extracting your own expertise and turning it into working systems. Each stage produces a concrete deliverable. The stages build on each other.

Surface

Make tacit expertise visible through structured elicitation.

Scope

Pick the expertise that codifies cleanly and produces the highest leverage.

Decompose

Break monolithic expertise into discrete operational components.

Specify

Write the intent for each capability with engineering-grade precision.

Build

Implement the capabilities as working AI systems using tools you choose.

Evaluate

Verify that the codified expertise produces correct outputs.

Operationalize

Integrate the systems into actual workflow so they do work you used to do.

Demonstrate

Produce a portfolio that proves the codified expertise to others.

Each stage takes roughly 5–10 hours of focused work. Most practitioners complete the program in 8–12 weeks.

Outcomes

Five outcomes, each tangible.

A portfolio of working AI systems that encode your expertise.

Not generic agents. Agents that do what you do — your code review patterns, your architecture evaluation heuristics, your debugging methodology, your domain knowledge. You own them. You keep them.

Demonstrable AI fluency at the level employers actually want.

Not "I used Copilot." Something more like: "I built a system that automates my team's PR review process using my own quality criteria, and it's running in production." The kind of artifact that ends performance-review anxiety and changes interview conversations.

Productivity multiplication that compounds.

Once your expertise is codified, you're directing work rather than doing it. Your throughput scales with infrastructure instead of hours. The 10–100x claim is honest for well-codified knowledge work because the AI is doing what you would do, just at machine speed.

A new professional position.

Your identity shifts from "experienced engineer" to "engineer who operationalizes expertise." That's a meaningfully different market position than the one most senior engineers occupy.

Permanent immunity to AI displacement fear.

Once you understand AI as an amplifier of your specific expertise, the fear dissolves. You're not racing the AI. You're directing it.

Fit

Who this is for. Who it isn't.

This is for you if…

→You have at least ten years of professional engineering experience.
→Your work involves significant judgment beyond following established patterns.
→You can articulate decisions you make that other engineers couldn't easily replicate.
→You're willing to invest 8–12 weeks at 5–10 hours per week in the methodology.
→You want a methodology, not a prompt library.

This isn't for you if…

—You have fewer than ten years of experience.
—Your work is primarily executing established patterns under direction.
—You're looking for quick wins, easy buttons, or AI shortcuts.
—You expect outcome guarantees in exchange for participation alone.
—You want to learn a specific tool rather than build a methodology practice.

The Program

Three ways to participate.

Self-Paced

The full methodology, on your schedule. Pre-recorded curriculum, deliverable reviews at stage gates, community access, monthly group office hours.

Founding cohort: $997

$1,997

Standard: $1,997

→Full methodology curriculum
→All templates, rubrics, and worked examples
→Stage-gate deliverable reviews (async)
→Cohort community channel
→Monthly group office hours
→12-month access to all materials

Enroll →

Live CohortFEATURED

Six to eight weeks of structured live work with a group of senior engineers. Weekly sessions, peer review, capstone presentation.

Founding cohort: $1,997

$3,497

Standard: $3,497

→Everything in Self-Paced
→6–8 weekly live sessions
→Small-group peer review breakouts
→Direct facilitator feedback
→Group capstone presentations
→Cohort-specific community channel

Apply →

Intensive (1:1)

Direct individual work on a specific high-stakes expertise area. Bounded engagement, weekend-only sessions, fixed deliverable set.

$9,997

→Everything in Self-Paced
→4–6 individual sessions (60–90 minutes each)
→Direct work on your specific expertise area
→Custom deliverable reviews with detailed feedback
→Pre-engagement scoping
→Post-engagement written summary and roadmap

Apply →

The founding cohort runs once. Pricing returns to standard after.

The Guide

Three decades of codifying expertise.

The Codified Practitioner methodology is built on three decades of work across named companies — Smashing Boxes, where I led the blockchain practice; Alekto, the AI company I founded; and earlier engagements across regulated industries. It draws on independent study (MIT Professional Education in data science, machine learning, and generative AI), public-domain research in knowledge engineering, and the practice of extracting expertise from senior engineers, including my own. Patent US8401960. Public writing on Intent Engineering. Currently leading software engineering at Microsoft Cloud + AI.

Full CV →

Questions

Common questions.

How much time does this require?

Plan for 5–10 hours per week over 8–12 weeks. The methodology is real work — there's no shortcut. The self-paced program is flexible on scheduling; the live cohort runs to a fixed calendar. Both expect the same total time investment.

Do I need to know AI tools already?

No. The methodology teaches you to apply AI to your expertise; you don't need to be an AI expert going in. We use Claude, OpenAI, and open-source frameworks in worked examples. If you can write code in any language, you have enough foundation.

What tools will I be using?

Claude (Anthropic) and OpenAI as primary platforms. Open-source agent frameworks where they fit. You bring your own API access. The methodology is tool-agnostic in principle, so you can apply it with whatever stack you prefer.

What if my employer cares about confidentiality?

Your codified expertise is yours. The methodology doesn't require sharing proprietary information — you work on expertise areas you choose, at whatever level of abstraction you're comfortable with. Several practitioners codify general engineering expertise rather than domain-specific knowledge that touches sensitive work.

What's the refund policy?

14-day refund window from purchase, before you submit your Stage 3 deliverable. Once you've completed the decomposition stage, you've consumed enough of the methodology that a refund isn't appropriate.

Do you guarantee outcomes?

No. The methodology produces outcomes for people who do the work; we don't guarantee that people will do the work. What we guarantee is that the methodology is real, the curriculum is rigorous, and the support during the program is direct.

Can I expense this through my employer's L&D budget?

Most practitioners do. The price points align with typical professional development expenditures. We provide an invoice on request that includes the program name and dates for reimbursement purposes.

Next Step

Apply for the founding cohort.

The founding cohort runs once. It's discounted by forty to fifty percent from standard pricing. The discount exists in exchange for testimonials and case studies that future cohorts will use as proof. Applications are reviewed individually. We're looking for senior practitioners with real expertise to codify — fit matters more than seniority alone.

Apply →