AI Coding for Real Engineers
A two-week cohort for developers who want to use AI coding tools on real production-style work: context gathering, planning, steering, feedback loops, AFK agents, and human-in-the-loop review.
With Max Techera — ex-CTO in Silicon Valley and AI Development professor at Universidad ORT.
2-week cohort · US$ 995
Limited seats · Live office hours · Recordings included
Most devs start AI coding the wrong way
There are two mistakes I see every day, and both hold you back. The way out isn't in either extreme.
Mistake #1 · Delegate everything
The agent writes spaghetti, you accept whatever, and in three weeks nobody understands the codebase. Speed today, tech debt tomorrow.
Mistake #2 · Delegate nothing
You keep doing it all by hand out of distrust. You burn out, ship slow, and leave the tool's whole productivity on the table.
The way out · The engineer's path
You steer the agent with judgment: context, a phased plan, and the loop closed with tests and review. You neither give up nor give up control.
This is the proactive middle path. It's not a productivity trick, it's a way of working.
What the engineer's path gives you
- Faster development without losing control of the code
- Better automated tests at the boundaries that matter
- Higher-quality software you can actually maintain
- More fun and less grind babysitting the agent
The engineer's path
First: analyze like an engineer
Engineering fundamentals matter MORE than before. The stakes: data loss, breaking changes, and losing understanding of your own code.
Second: think like an engineer
Communicate, anticipate, plan, decompose, delegate, systematize, parallelize. The same skills you use with humans are the ones you use to steer agents.
The 7 skills you'll practice
Steering an agent is steering an engineer. These are the competencies we work on, in your own code, across the 2 weeks.
Communicate. Give the agent exactly the context it needs to do the right thing.
Anticipate. See where it'll fail before it writes a single line.
Plan. PRDs and multi-phase plans instead of one giant prompt.
Decompose. Split big features into tracer bullets you validate fast.
Delegate. Know what to hand off and what to review yourself.
Systematize. Skills, AGENTS.md, and conventions that repeat themselves.
Parallelize. Run AFK agents without losing control or quality.
Why I built this cohort
I'm not an AI guru. I'm a CTO and I write code with Claude Code every day, in production, under the same constraints you have: deadlines, code review, and a codebase that can't break.
What changed for me wasn't delegating more. It was learning to STEER the agent the way you steer an engineer: giving it context, planning in phases, and closing the loop with tests and review. That's what I teach here, the real system, not productivity tricks.
The cohort is live and deliberately small: I want to review your code and answer you directly, not run a generic masterclass.
Vibe coder vs. Real Engineer
“Went from not understanding Claude Code to automating my complete workflow in a week. The sub-agents content is pure gold.”
Is it for you?
Yes, if...
- You already code and want to use AI on real work, not demos.
- You're tired of the spaghetti code agents spit out without direction.
- You want a repeatable system, not scattered productivity tricks.
- You can spend a few hours a week for 2 weeks.
No, if...
- You're looking for a “learn to code from scratch” course.
- You want AI to do everything without understanding what it does.
- You're not willing to read code or review what the agent generates.
The curriculum
Getting to know Claude Code
- Setup and fundamentals
- Your first Plan / Execute / Clear loops
Fundamentals, steering & planning
- Context window management and the Plan/Execute/Clear loop
- Steering with AGENTS.md and custom skills
- Planning with PRDs and tracer bullets
- Multi-phase feature decomposition
- Live office hours
Feedback loops, AFK agents & human-in-the-loop
- CI pipeline and quality testing practices
- The "Ralph Wiggum" loop for autonomous building
- Human-in-the-loop vs. fully autonomous modes
- Research phases and kanban workflows
- Live office hours
Format & access
Built so you can keep up even with a full-time job.
- New lessons every Monday, self-paced, no fixed schedule
- 6 live office hours across 3 days and multiple timezones
- Everything recorded with transcripts, catch up whenever you want
- Private Discord community during and after the cohort
- Lifetime access to the lessons, code, and recordings
What's included
- Video lessons + hands-on exercises
- Course code repository
- Pre-course access
- 6 live office hours (3 days, multiple timezones)
- Lifetime recordings and transcripts
- Private Discord community
Max Techera
From Montevideo to Silicon Valley. I wasn't born into the tech ecosystem, I coded my way in.
15+ years building apps that hit #1 on the App Store. As Director of Product Engineering at LastRev I led architecture for enterprise clients; as CTO of AnswerAI I built an AI platform from scratch.
Today I teach AI Development at Universidad ORT and make content for the Spanish-speaking developer community.
What I teach, I use. Every. Single. Day.
30-day guarantee
If you do the work (attend office hours, complete the exercises) and feel it wasn't worth your investment, I refund 100%.
No questions, no drama. I offer this because I know what the program delivers: if you do the work, the results follow.