AI Engineering Cheatsheet

The new tools of the trade

You Are Here

You are here because you can just do things. You were selected because you’re agentic, a self-starter, and have the core traits needed in today’s AI-empowered world: systematic thinking, strong communication skills, and a bias toward making messy things legible

Here are the new tools of the trade

The way we work has fundamentally changed. AI means that structured problem-solvers — the people trained to break down ambiguous problems, communicate clearly, and orchestrate complex workstreams — can now build production software. The technical bar has dropped. The thinking bar hasn't. You already have what it takes

This archetype — agentic, systematic, communication-first — has real potential to outperform traditional software engineers in the AI era. A pure technician can write code. You can hold the whole problem in your head. The ability to decompose ambiguity, manage stakeholders, and context-switch across domains is exactly what modern AI-assisted engineering rewards. When the bottleneck shifts from typing code to knowing what to build and why, the consultant-turned-engineer has a real edge

“Generalists who straddle disciplines — business, design, infrastructure, and user needs all at once.”

Boris Cherny, Head of Claude Code — on the profile of the new builder

Vibe coding creates maintenance debt faster than it creates value. Everyone can code now, but that doesn't mean everyone knows how to build. The difference is understanding the layers underneath — what makes something shippable, maintainable, and real


What This Is

Two modules. The minimum you need to go from “I've never deployed anything” to shipping a working application into production

Level Up — The Curriculum

The ten layers of shipping something real. Environments, version control, databases, hosting, domains, auth, scheduled jobs, workers, scale, and system design — each explained as a standalone concept with the context you need to understand why it matters

The Stack — Modules in Practice

What reusable modules look like and why you should build them. Examples of packaging common patterns — notifications, data feeds, scraping — so every future project starts from a higher baseline

Start with Level Up. That’s the foundation — the layers every production app needs regardless of what you’re building. The Stack shows you what to build on top once you have the fundamentals


The Engagement Model

Think of this like a consulting deck. The framing page is what you need to ship. The deep dives are each layer unpacked

LayerWhat it doesOptions
ApplicationThe code that solves a problemPython, React, Flask
Version ControlTrack changes, collaborate, undo mistakesGit, GitHub
DatabasePersist data between sessionsSupabase, Postgres, SQLite
SearchFind things in large datasetsTF-IDF, cosine similarity, pgvector
HostingRun your code on the internetVercel, Dokku, Railway
RoutingDomain, DNS, SSLCloudflare, Squarespace DNS
AuthControl who accesses whatOAuth, JWT, session cookies
Scheduled JobsRun code when nobody is watchingCron, Celery Beat, GitHub Actions
Workers & ScaleMove slow work out of the request pathCelery, RQ, nginx, HAProxy
System DesignStructure ambiguous technical problemsRequirements, trade-offs, MVP vs. scale

Each of these is a chapter in Level Up. Master them and you can ship anything


The Long Game

You don’t need to master everything before you start building. You need to understand the layers, then go find yourself a problem to solve. Every engagement sharpens the next one

Once you have the foundation, you start building reusable modules — The Stack shows what that looks like. Notifications, data feeds, scraping, inference — each one packaged so it can be imported into any future project with one line

The structured problem-solving, communication, and orchestration that made you a good consultant will make you the ideal forward deployed engineer

Get in touch