05 Reference 10 min read 2,031 words
π Recommended AI Courses & Learning Paths
A curated list of reliable, trusted, and up-to-date online courses for AI engineers, ML practitioners, and product teams. Every course here is verified as of March 2026 β no fluff, no outdated MOOCs.
referencecareersintro
Foundation: LLMs & Transformers
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Neural Networks: Zero to Hero | Andrej Karpathy (YouTube) | Free | The definitive from-scratch series by an OpenAI/Tesla legend. Builds GPT from scratch. |
| CS324: Large Language Models | Stanford | Free | Stanford-quality lecture notes covering LLM fundamentals, scaling laws, alignment. |
| Generative AI with LLMs | DeepLearning.AI + AWS (Coursera) | ~$50 | Hands-on intro to LLMs, covering training, fine-tuning, RLHF. By Andrew Ng's team. |
| Practical Deep Learning for Coders | fast.ai | Free | Bottom-up, code-first approach. Best for engineers who learn by doing. |
RAG Pipelines
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Building and Evaluating Advanced RAG | DeepLearning.AI + LlamaIndex | Free | Covers sentence-window retrieval, auto-merging, RAG evaluation with TruLens. |
| Vector Databases: from Embeddings to Applications | DeepLearning.AI + Weaviate | Free | Practical walkthrough of embeddings, vector stores, and hybrid search. |
| Building RAG Agents with LLMs | NVIDIA Deep Learning Institute | Free | Enterprise-grade RAG with NVIDIA NIM. Covers chunking, reranking, evaluation. |
| LlamaIndex β Documentation: Learning | LlamaIndex | Free | Official LlamaIndex learning path β best for deep RAG pipeline mastery. |
| RAG Fundamentals (Haystack) | deepset / Haystack | Free | Hands-on tutorials for pipeline-based RAG using the Haystack framework. |
Agentic AI & Multi-Agent Systems
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| AI Agents in LangGraph | DeepLearning.AI + LangChain | Free | LangGraph by the creators. Covers ReAct, persistence, human-in-the-loop, multi-agent. |
| Multi AI Agent Systems with crewAI | DeepLearning.AI + crewAI | Free | Official CrewAI course. Covers Crews, Flows, and real-world business automations. |
| Building Agentic RAG with LlamaIndex | DeepLearning.AI + LlamaIndex | Free | Routing, tool-calling agents, and multi-document agentic retrieval. |
| Functions, Tools and Agents with LangChain | DeepLearning.AI + LangChain | Free | Tool-calling, OpenAI function calling, building from scratch. |
| Developing AI Agents using AutoGen | DeepLearning.AI + Microsoft | Free | AutoGen multi-agent patterns. Covers debate, tool-use, and code execution agents. |
| CS294/194-196: LLM Agents (Berkeley) | UC Berkeley | Free | Graduate-level course on LLM agents. Covers memory, planning, safety, evaluation. |
Context & Memory Management
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Building Systems with the ChatGPT API | DeepLearning.AI + OpenAI | Free | Covers multi-turn conversation state, context management, moderation chains. |
| Prompt Engineering with Llama 2 | DeepLearning.AI + Meta | Free | Shows context window tradeoffs and system prompt management with Llama 2. |
| Reasoning with o1 | DeepLearning.AI + OpenAI | Free | Deep dive into o1 reasoning, budget tokens, thinking modes. Directly applicable to o3 and Claude 3.7 Extended Thinking. |
| Mem0 Documentation | Mem0 (official) | Free | Definitive reference for multi-tier memory in production agents. |
AI Evaluations & Observability
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Evals for AI: Maven Course | Hamel Husain & Shreya Shankar (Maven) | Paid (~$400) | Industry gold standard for AI evals. Used in production at dozens of companies. Our repo's evals guides are based on this course. |
| Evaluating and Debugging Generative AI | DeepLearning.AI + W&B | Free | Covers tracing, evaluation with W&B Weave, and experiment tracking. |
| Quality and Safety for LLM Applications | DeepLearning.AI + WhyLabs | Free | Covers hallucination detection, toxicity, bias evaluation, and drift monitoring. |
| LangSmith Evaluation Tutorials | LangChain | Free | Official LangSmith docs are the best hands-on eval reference if you use the LangChain ecosystem. |
| Phoenix + Langfuse official docs | Arize Phoenix | Free | Hands-on tutorials for open-source evals using Phoenix. |
Prompt Engineering & Context Engineering
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| ChatGPT Prompt Engineering for Developers | DeepLearning.AI + OpenAI | Free | The foundational prompt engineering course. By Isa Fulford & Andrew Ng. |
| Prompting Fundamentals (Anthropic) | Anthropic | Free | Straight from the Claude team. Covers prompt design, XML tags, chain-of-thought. |
| DSPy: Building Optimizable Pipelines | Stanford NLP (GitHub) | Free | Not a course but the DSPy repo's notebooks are the best way to learn programmatic prompting. |
| Prompt Engineering Guide | DAIR.AI | Free | Comprehensive, community-maintained reference covering all major prompting techniques. |
Fine-tuning & Adaptation
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Finetuning Large Language Models | DeepLearning.AI + Lamini | Free | Covers LoRA, full fine-tuning, dataset preparation, evaluation. Concise and practical. |
| Reinforcement Learning from Human Feedback | DeepLearning.AI + AWS | Free | Deep dive into RLHF: reward models, PPO, preference datasets. |
| Hugging Face NLP Course | Hugging Face | Free | The best free course for fine-tuning transformers with the HF ecosystem (Trainer, PEFT, etc). |
| How Diffusion Models Work | DeepLearning.AI | Free | For image model fine-tuning (stable diffusion, LoRA for images). |
Inference Optimization & MLOps
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| ML Engineering for Production (MLOps) | DeepLearning.AI (Coursera) | Paid | 4-course specialization on production ML: deployment, monitoring, pipelines. |
| Efficiently Serving LLMs | DeepLearning.AI + Predibase | Free | Covers vLLM, PagedAttention, quantization, LoRA serving. Exactly what this guide covers. |
| vLLM Documentation & Tutorial | vLLM | Free | The official vLLM docs are the most up-to-date reference for high-throughput serving. |
AI Safety & Guardrails
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Red Teaming LLM Applications | DeepLearning.AI + Giskard | Free | Hands-on red teaming, prompt injection, jailbreak detection, bias testing. |
| AI Safety Fundamentals | BlueDot Impact | Free | Most trusted free course on AI alignment and safety. Used by professionals at Anthropic, DeepMind. |
| NVIDIA AI Red Team (NEMO Guardrails) | NVIDIA | Free | Hands-on notebooks for building production guardrails with NeMo Guardrails. |
Coding Agents & Developer AI Tools
| Course | Provider | Cost | Why It's Trusted |
|---|---|---|---|
| Claude Code β Official Docs | Anthropic | Free | The definitive starting point for Claude Code. Covers CLAUDE.md, SDK, and permissions. |
| Building Code Agents (Hugging Face) | Hugging Face | Free | HuggingFace's official agents course β includes a unit on building code-execution agents. |
| Introduction to OpenHands | All-Hands AI | Free | Official getting-started guide for OpenHands autonomous coding agent. |
For PMs & Non-Engineers
These require no Python experience:
| Course | Provider | Cost | Why It's Good |
|---|---|---|---|
| AI for Everyone | DeepLearning.AI (Coursera) | Free | Andrew Ng's course for non-technical roles. Covers what AI can/can't do, project leadership. |
| Prompt Engineering for Everyone | Learn Prompting | Free | Plain-English guide to prompt engineering for non-engineers. |
| Evals for AI (Maven) | Hamel Husain & Shreya Shankar | Paid | Despite having code, this course is designed for PMs and QAs, not just engineers. Highly recommended. |
| AI Product Management | Product School | Free (blog) | Practical guide for PMs building AI-powered products. |
| Google: Introduction to Generative AI | Google Cloud Skills Boost | Free | No-code introduction to generative AI, LLMs, and responsible AI. |
YouTube Channels & Free Content
| Channel / Resource | Focus | Why Follow |
|---|---|---|
| Andrej Karpathy | Foundations, transformers | Best explanations of how LLMs actually work |
| Yannic Kilcher | Paper reviews | Clear walkthroughs of latest ML research papers |
| Aleksa GordiΔ - The AI Epiphany | Paper reviews | Deep technical paper breakdowns |
| AI Jason | Agents, LangChain, practical | Great intro videos for agentic frameworks |
| Sam Witteveen | Gemini, RAG, agents | One of the best practical AI YouTubers |
| Matt Wolfe | AI news, product demos | Best for staying current on AI news and tools |
| Hamel Husain (blog) | Evals, production AI, LLMs | Real production insights from the author of the evals maven course |
| Simon Willison (blog) | LLM news, tools, coding | The most trustworthy daily AI news source |
| The Latent Space podcast | Technical AI interviews | Best technical AI podcast β deep dives with researchers |
| Lex Fridman Podcast | Broad AI/ML interviews | Long-form interviews with leading AI researchers |
Learning Path Suggestions
13.1
π€οΈ Path: "I'm new to AI and want to build things fast"
12345
Week 1: Prompt Engineering for Developers (DeepLearning.AI) β free, 2 hrs
Week 2: Building Systems with ChatGPT API (DeepLearning.AI) β free, 2 hrs
Week 3: Building and Evaluating Advanced RAG (DeepLearning.AI) β free, 2 hrs
Week 4: AI Agents in LangGraph (DeepLearning.AI) β free, 4 hrs
Month 2: Pick a real project, use this guide as reference
13.2
π€οΈ Path: "I want to understand LLMs deeply"
1234
Week 1-3: Neural Networks: Zero to Hero (Karpathy) β free, 12+ hrs
Week 4-6: CS324 Stanford LLMs β free, 30+ hrs
Month 2: Generative AI with LLMs (Coursera DeepLearning.AI)
Month 3: CS294 LLM Agents (Berkeley)
13.3
π€οΈ Path: "I want to build production-ready AI evaluation"
1234
Week 1: Evaluating and Debugging Generative AI (DeepLearning.AI + W&B) β free
Week 2: This repo's evals guides (Phoenix/Langfuse) β free β start here
Week 3-4: Quality and Safety for LLM Applications (DeepLearning.AI)
Month 2: Evals for AI (Maven, Hamel + Shreya) β paid, worth it
13.4
π€οΈ Path: "I'm a PM learning to contribute to AI product quality"
1234
Week 1: AI for Everyone (Coursera) β free
Week 2: Prompt Engineering for Everyone (learnprompting.org) β free
Week 3: AI Evals guide in this repo β free (especially Chapters 1-3 on error analysis)
Month 2: Evals for AI (Maven) β paid, has PM track
13.5
π€οΈ Path: "I want to deploy coding agents in my team"
1234
Day 1: Claude Code docs (anthropic.com) β free
Week 1: This repo's Claude Code Guide + OpenCoder Landscape Guide
Week 2: Building Code Agents (Hugging Face) β free
Month 1: Run Claude Code on a real project in CI
How to Stay Current
AI moves fast. Beyond courses, these habits keep you current:
- Follow Simon Willison's blog β daily, trustworthy AI news summaries
- Read Anthropic + OpenAI release notes β primary sources beat second-hand summaries
- Watch the Latent Space podcast β best technical depth
- Contribute to open source β OpenHands, LlamaIndex, DSPy β real learning happens in PRs
- Star this repo β we update it as the landscape changes β
Maintained by Om Bharatiya. PRs welcome for new course additions!
Key takeaways
01
Almost the entire list is free
Only three courses carry a price: the Maven evals course at about $400, Coursera's Generative AI with LLMs at about $50, and the MLOps specialization. Everything else costs nothing.
02
Five paths, chosen by intent
The tracks are build fast, understand LLMs deeply, build evaluation, contribute as a PM, and deploy coding agents, each given a week-by-week schedule rather than a ranked list.
03
The depth track runs three months
Understanding LLMs deeply budgets 12+ hours for Karpathy's Zero to Hero and 30+ hours for Stanford CS324 before reaching Berkeley's CS294 agents course in month three.
04
Blogs and podcasts count as sources
Simon Willison's blog, Hamel Husain's blog and the Latent Space podcast sit alongside the courses, and the closing section argues for vendor release notes over second-hand summaries.