About aiml
AI engineering courses with runnable code
aiml.sh is Arjun Prabhulal’s site for learning LLMs, AI agents, retrieval and evals with working code. Follow an explanation, inspect the implementation and test where it fails.
Choose your starting point
Each course teaches one part of an AI application in order: what you'll learn, its duration and prerequisites, then the lessons.
Start
Start with AI Fundamentals
Run a local model, work with tokens and embeddings, validate tool calls and write a first eval.
Start AI FundamentalsImplement
Already write Python?
Build an agent with Google ADK or LangGraph, then connect tools through MCP and add retrieval and evals.
Start with Google ADKNext
Follow the learning path
Start with AI Fundamentals, choose an agent framework, then add MCP, RAG and LLM Evals. Optional ADK extensions cover runtime controls, tracing, guardrails and live audio.
View the learning pathWhat to expect from a lesson lab
Each lab starts from a concrete problem and an input/output contract. The code, diagrams and examples describe the same implementation.
- Reproduce the behavior
- Use the supplied code, setup commands, inputs and expected output. Change an input and compare what happens with your prediction.
- Inspect the failures
- Run the failure tests and follow the primary references. Each measurement applies to the hardware and workload stated with it.
Read and practice at your pace
Courses, lessons and labs are free. Running an example may require your own hardware or an external service; check its requirements first.
Saved and completed lessons stay in this browser and do not sync between devices. Find them in My learning.
Related sites
whatisai.fyi provides short definitions and sourced explanations. aiml focuses on implementation, system boundaries and failure behavior. Follow new lessons via RSS.
Who writes aiml.sh
Arjun Prabhulal writes aiml.sh. He also publishes hands-on guides on Google ADK, MCP and agent systems at arjunprabhulal.com. Selected guides: