Data-driven field guide to AI engineering roles, skills, and interviews. AI Engineering Field Guide is the go-to resource for AI engineering.
Everything here is based on real data: 4,894 actual job descriptions, real interview experiences, and real stories from practitioners. This is not AI-generated filler dumped into a repo - every insight comes from analyzing actual data and synthesizing patterns from it.
It's a work in progress, and I'm actively adding more content. Your input is very welcome - feedback and contributions help shape what goes in here.
Star this repo to keep an eye on updates. To get notified about new content, subscribe to my newsletter: Alexey on Data.
- My vision of the role - how I see AI engineering, comparison with DS/ML/DE roles, CRISP-DM for AI
- Skills analysis - top skills, job types, cloud platforms, frameworks
- Responsibilities - patterns extracted from 5,694+ job responsibilities
- Use cases - 4,525 real use cases showing what companies build with AI
- Reality vs. job postings - what candidates experience vs. what's advertised
- Interview process - common patterns, step counts, time estimates, AI use in hiring, key takeaways
- Interview questions - consolidated from 100+ sources
- Theory - LLMs, RAG, agents, ML fundamentals, company-specific questions
- Coding - coding round formats, DSA problems, ML implementation exercises
- Project deep dive - presentation rounds, follow-up probes, what interviewers evaluate
- AI system design - system design for AI applications
- Behavioral - values, leadership, problem-solving
- Home assignments - take-home assignments and paid work trials from 100+ GitHub repos
- Skills that get you hired - baseline expectations, differentiators, and portfolio strategy
- After the interview - handling offers, rejections, and salary negotiation
- Interview trends - realistic assessments, AI cheating, AI-proctored rounds
- Company-by-company data - individual interview process descriptions for 51 companies, linked to source job postings
- General learning path - what to learn and in what order
- From Data Engineer - smoothest transition, 3-4 months
- From Data Scientist - evaluation is your superpower, add engineering
- From ML Engineer - easiest transition, replace model call with API call
- From Backend Engineer - 2-3 months, add AI on top of engineering
- From Frontend Engineer - backend first, then AI, unique full-stack advantage
- Project ideas - real project examples that demonstrate AI engineering skills
4,894 job descriptions scraped from builtin.com covering LA, NY, London, Amsterdam, Berlin, and India.
- Structured job descriptions - YAML files grouped by scrape date
- Raw extracted postings - original extracted data grouped by scrape date
Curated collection of resources we compiled while researching content for this field guide:
- Practitioner interview stories
- AI system design guides
- Company engineering blogs
- Books and courses
- Case study collections
See awesome.md for the list.
A 4-part event series on AI engineering careers, hosted through Maven and AI Shipping Labs:
- A Day of an AI Engineer - the practical reality of the role (Maven, AI Shipping Labs) - recording available
- Defining the AI Engineer Role - what companies actually hire for, based on 2,400+ job descriptions (Maven) - recording available
- The Interview Process - real hiring trends, technical questions, and live coding challenges (Maven) - recording available
- Take-Home Assignments - analyzing real assignments and building production-ready solutions (Maven) - recording available
There's more to learning AI engineering than watching tutorials:
- AI Engineering Buildcamp: From RAG to Agents - my 9-week intensive course on building production-ready AI applications, covering RAG, agents, testing, evaluation, and monitoring
- AI Shipping Labs - a community of practitioners building AI and with AI, with workshops, courses, case studies, and discussions on what actually works