Modern LLM Notebook
WalkingLabsA free notebook course that builds an LLM stack by hand, from BPE and GPT to MoE, DPO, speculative decoding and vLLM serving.
A curated library of the best AI learning resources—courses, tutorials, docs, and talks on agentic AI, MCP, Agent Skills, RAG, machine learning, fine-tuning, models, and frameworks, for AI engineers and technical teams.
A free notebook course that builds an LLM stack by hand, from BPE and GPT to MoE, DPO, speculative decoding and vLLM serving.
How LLMs encode token position, from GPT-2's learned table to RoPE, partial RoPE and NoPE layers, with PyTorch code.
Anthropic's method for designing, grading and reporting LLM safety evals, with sample-size floors and grader checks you can copy.
How to design an eval you can trust and improve against it without overfitting, using claude-api build-eval and hillclimb.
When one agent loop with skills and tool search beats a router plus specialist subagents, and how to migrate in stages.
The five primitives behind a stateful, proactive agent: event inbox, timers, memory, permissions and outbox.
The resources learners opened most this week.
The open framework Prime Intellect used to post-train INTELLECT-3, with docs on async RL, loss design and scaling to large GPU clusters.
Pull structured entities out of long documents with an LLM and trace every one back to its exact character offsets.
Read a 100-line coding agent that scores above 74% on SWE-bench Verified, then build your own in about 60 lines of Python.
LangChain's free course on taking a deep agent to production on LangSmith's managed runtime, with sandboxes, connections, schedules, auth, channels and evals.
A free Microsoft course that takes JavaScript and TypeScript developers from a first LLM call to agentic RAG and MCP servers with LangChain.js.
A project-driven Python book that goes from your first programmatic prompt to tool-using LangGraph agents, advanced RAG and your own MCP servers.
A four-module CrewAI course that takes one code-review agent from prototype to a monitored, CI/CD-deployed multi-agent system.
Use TabPFN to predict on spreadsheets without training a model, then hand it to an LLM agent as a tool.
Generate validated synthetic training and evaluation data from samplers, LLMs, seed datasets, and agent traces.
Build agent environments and verifiers, run repeatable benchmarks, and feed the same setup into RL training.
Follow a large language model from GPU performance limits through pre-training, post-training, and production inference.
A seven-week master's course with public slides and labs that takes you from a tiny GPT to a tool-using agent.
Build Qwen3 inference, a mini vLLM and a coding agent from raw MLX array operations on a Mac.
Validate LLM output, user input and tool arguments with Pydantic models in a customer support assistant.
Use Google's ADK to build agents that turn CSV files and product reviews into one Neo4j knowledge graph.
Build an arXiv research assistant in seven stages, from Docker services and BM25 to hybrid search, tracing and a LangGraph agent.
The official reference for Kiro's spec-driven agent workflow: requirements, design and task files, steering, hooks, MCP and custom agents across IDE, CLI and web.
Anthropic's free developer course for building Claude apps on Amazon Bedrock with boto3, from prompt evals and tool use to RAG, MCP and agents.