AI for backend
engineers.
Practical guides for engineers who can ship a REST API but haven't built with AI yet. From your first LLM call to production RAG, MCP servers, and evals. Free, no paywalls. Tips welcome.
The topics
Each topic is a structured curriculum, not a blog. Built so you can go from zero to shipping production code, one article at a time.
AI Engineering
The core curriculum. Foundations, RAG, evals, production patterns. From "I've never called an LLM API" to "I shipped this to prod."
MCP Development
Build with the Model Context Protocol. Servers, clients, transports, integrations.
AI Agents
Build agent systems backend engineers can trust. Tool use, planning, state, retries, durable execution.
AI Coding Tools
Master Cursor, Claude Code, Windsurf. Agentic workflows for senior engineers.
LLM Infrastructure
Production LLM serving. vLLM, batching, GPU economics, autoscaling, caching. The infra behind AI at scale.
Voice & Real-time AI
Low-latency voice agents. STT, TTS, streaming, interruption handling, WebSockets. The backend behind real-time AI.
LLM Fine-Tuning
LoRA, dataset prep, eval-driven tuning. From notebook to production model.
How it works
Start with AI Engineering
The core curriculum is the first live topic. Read in order if you are new to AI, or jump to the article that matches what you are building.
Read what you need
Read in order if you are new to AI, or jump to the topic you are building right now.
New articles when they ship
Each article is ~10-15 minutes. Diagrams, examples, no fluff. Sent to your inbox when it ships, or read on the site.
Get new articles by email
Optional. I send an email each time a new article goes up, no more often than that. No spam, no upsells, unsubscribe in one click.
Powered by Buttondown. Your email is only used to send articles.
Buy me a coffee
Everything on AI Engineer Path is free, forever. No paywalls. If a guide saved you a debugging weekend or helped you ship a feature, throw a few dollars in the jar. It keeps me writing.
- Nothing changes for you, the content stays free
- I get a little caffeine for the next article
- Add a note at checkout if you want; I read all of them
- Roughly the cost of a sip of coffee
- Recurring, but cancel anytime
- Helps me keep the site ad-free, forever
- Same price as monthly, less paperwork
- One less recurring bill on your statement
- Cancel anytime, prorated refund
Who this is for
Backend engineers who've shipped production services and now want to add AI to their toolkit, starting from zero.
Engineers who got an LLM call working and now want to do it well. Caching, evals, observability, the production stuff.
Senior engineers who want depth, not tutorials. Explanations grounded in real code, with the decisions you'll actually face.
Frequently asked questions
Is this really free?
Yes. Every article is free, forever. No paywalls, no signup walls, no email harvesting. If you find the content useful and want to support it, the "Buy me a coffee" section has three optional tiers. None of them unlock anything; they just keep the site going.
What do I get if I donate?
Nothing extra on the site. The same articles, same diagrams, same code samples. What you get is the warm feeling of supporting independent writing and helping me say no to ads forever. If you add a note at checkout, I read it and write back.
What topics will you add?
AI Engineering (live) and MCP Development (live) are the foundation. Next planned: AI Coding Tools (Cursor, Claude Code, Windsurf workflows), LLM Fine-Tuning (LoRA, datasets), AI Agents (tool use, planning, durable execution), LLM Infrastructure (vLLM, batching, GPU economics), and Voice & Real-time AI (low-latency voice agents). New topics ship when they are ready.
I'm new to AI development. Is this for me?
Yes. Every article is written for someone who codes but is new to LLMs, RAG, embeddings, MCP, etc. Technical terms are defined when introduced. Start with the first article in either topic.
I'm a senior engineer. Will I learn anything?
The articles are accessible but the depth is real. Each one ends with the production decisions you'll face and the trade-offs that matter. If you've shipped AI in production for years you'll skim some sections, but most readers find at least 2-3 new patterns per article.
How often do new articles ship?
There is no fixed schedule. Articles ship when they are ready, typically a few per month.
Can I cancel my recurring support?
Yes, anytime. Lemon Squeezy sends a cancel link in every receipt, or you can email and I will cancel it for you within a day. Monthly stops at the next billing date. Annual gets a prorated refund for unused months. No questions.
Who runs this site?
Working engineers who've built and shipped AI in production at scale. No LLM-generated content, ever. See the About section below.
About AI Engineer Path
AI Engineer Path is a learning resource for backend engineers building with AI. Written by people who've shipped production AI, not consultants summarizing papers. Beginner-friendly to start; deep enough to take you to production.
The goal: write the resource I wish existed when I went from "I can ship a REST API" to "I just put an AI feature in front of real users." Clear explanations. Honest about what is hard. Free, because price should not be the reason anyone goes without it.