Engineering Coaching

AI is an accelerator, not a compass


Projects rarely stall at the start. They stall at real complexity, and AI will happily keep adding code past the point where that helps. The code still counts, but the decisions behind it decide whether what you build holds up.

That judgment isn't something you can prompt your way to. I'm Bob, and I coach it on your own code, one on one or in a small cohort, on real projects with real stakes. I meet you where you are and take you to the next level.

Sun/Oracle · 10+ years 150+ developers coached 1,000+ pull requests reviewed Co-founder Pybites

Read my full story →

A plan built for you, one real project, deep human code reviews

I meet you where you are

There's no fixed curriculum. I find out what you know and where your gaps are, and we work from there. Someone writing their first real app and someone with fifteen years in get a different program: same attention, different plan, both built around your goal.

Read student stories →

You build one real thing

Not exercises. A project with real edge cases, deployed and built for actual users. Developers I've coached ship things like a payroll SaaS with paying clients, a legal citation tool for Canadian courts, a distributed lock server over gRPC, and a pytest plugin published on PyPI.

See what they shipped →

Detailed code reviews

AI can review code, but it misses context and doesn't tailor its advice to your level and learning style. Doing that takes real attention, and students grow the most from the discussions that come out of it: not just what to change, but why it matters in their codebase.

Why I review every PR myself →

The three pillars of modern software engineering

Decision-making

Software engineering has always been about architectural choices. We're paid to solve business problems, and writing code is only part of that.

Read →

The control layer

LLMs are tools. The real work is the control layer around the model. Reliable software needs repeatable architectural patterns.

Read →

Design over code

Real productivity lives in the questions you ask before coding. In one case, replacing a 1,069-line AI-generated app with 156 lines came down to experience and design mastery.

Read →

What developers say after working with me

See all stories →

Mastery over hype

Get The Zen of Python in Code Review, free: 19 lessons on the judgment that turns AI speed into code that lasts, one line of import this at a time. Plus one email a month on Python, Rust, and AI.

Featured articles

June 4, 2026

"Rust Is for People Who Want to Be Punished." Now Jochen Trusts It More Than Python.

Jochen thought Rust was self-punishment for the pros. Three months later, his hand-rolled JSON parser ran up to 3.5x faster than CPython.

Read →
May 20, 2026

How to Keep Your Developer Instincts When AI Writes the Code

AI agents will write your code if you let them. Here's the deliberate practice routine I use to stay technical when AI increasingly writes code for me.

Read →
April 20, 2026

From Hobby Code to SaaS to Orchestrating AI Agents

Ryan Austin ships payroll software in the Bahamas. In 2021, he couldn't tell if his architecture was sound. In 2025, he's directing AI agents with confidence. The foundation made the difference.

Read →
All articles →

Become the engineer who owns the system.

Tell me what you're working on and what's standing in the way to reach your goal.

Prefer getting on a call to talk it through? Book a free call →

“Six years later, I can see that doing the coaching was a great decision. In January I was promoted to Senior AI Engineer and became a team coordinator.” — Piotr R., Senior AI Engineer at Lingaro