HomeAboutProjectsRadarUniverseSkillsAchievementsContact
🌙
Back to the radar

AI Engineer Roadmap 2027

instagramreview
Original post

TL;DR

A six-step self-study roadmap for becoming an AI engineer, curated by a senior AI engineer, pointing to free resources covering Python, ML fundamentals, LLMs, prompt/agent engineering, cloud, and evaluation. Aggregates free, structured learning resources across the full AI engineering stack—from Python fundamentals to agent frameworks and evaluation—reducing the effort of self-curating a curriculum.

What it actually is

  • What: A six-step self-study roadmap for becoming an AI engineer, curated by a senior AI engineer, pointing to free resources covering Python, ML fundamentals, LLMs, prompt/agent engineering, cloud, and evaluation.
  • Who built it / maintained by: Khalil (Instagram: khaliladib11), self-described senior AI engineer based in London
  • Status: unknown
  • Why it matters: Aggregates free, structured learning resources across the full AI engineering stack—from Python fundamentals to agent frameworks and evaluation—reducing the effort of self-curating a curriculum.
  • How it compares to alternatives:
  • roadmap.sh AI engineer path
  • fast.ai Practical Deep Learning
  • DeepLearning.AI specializations
  • Andrej Karpathy's Zero to Hero
  • GitHub stars: unknown · License: unknown · Archived: unknown

Links

  • (no links found)

Research evidence

  • Review limitation: Repository identity unresolved: use extracted repository links or matching archive identity, not a guessed repository

General descriptions are reported, not independently tested. source_verified means the exact quotation was found in retrieved source text; it does not establish that an upstream claim works locally.

  • Identity: unresolved · Basis: unresolved
  • reported: All six learning resources in the roadmap are free (evidence: none)
  • reported: The roadmap has six steps: Python, ML Fundamentals, LLM Basics, Prompt/Agent Engineering, Cloud, and Evaluation (evidence: none)
  • reported: Step 1 includes Python topics: types, async, Pydantic, FastAPI (evidence: none)
  • reported: Step 3 covers LLM basics including tokens, context windows, and structured output (evidence: none)
  • reported: Step 4 covers prompt/agent engineering including RAG, tool calling, MCP, and agents (evidence: none)
  • reported: Step 6 (evaluation) is described as the step most learners skip (evidence: none)
  • reported: Harvard CS50's Introduction to Programming with Python is recommended for step 1, with a 10-week course length of 3–9 hours per week (evidence: none)
  • reported: The Hugging Face LLM Course is recommended and is divided into three parts: LLM Fundamentals, LLM Scientist, and LLM Engineer (evidence: none)
  • reported: An AWS cloud training course by @techwithsoleyman is recommended for the cloud step (evidence: none)

Kickstarter guide

Pseudocode / untested proposal. Check current upstream instructions and the target project before execution.

This is a curated roadmap, not a software library. To follow the roadmap as described in the post, one would start with Harvard CS50's Introduction to Programming with Python (free on edX), then study ML fundamentals via a Stanford or Hugging Face course, then work through the Hugging Face LLM Course (huggingface.co/learn/llm-course), then complete a project-based agent/harness engineering course, then take an AWS cloud fundamentals course, and finally study AI evaluation practices. All steps are reported as free. No install commands or code imports apply.