LLMs & Agentic AI

Comprehensive guide to zero-shot, few-shot, chain-of-thought, ReAct, and other prompting techniques with practical chemistry-relevant examples.

Free courses on building LLM applications: RAG pipelines, agents, memory, and tool use with LangChain and LangGraph.

Prolific practitioner blog covering LLM capabilities, limitations, tool use, and agentic workflows. Consistently useful for staying current.

Step-by-step notebook for building a RAG pipeline: document loading, chunking, embedding, vector store, and retrieval-augmented QA.

Reference implementation of a chemistry agent with 18 tools: PubChem lookup, RDKit, reaction prediction, and safety checks. Directly relevant to Tutorial 14.

Introduces the ReAct framework interleaving reasoning traces and action calls. The foundation for tool-using LLM agents.

Foundational RAG paper combining a retrieval module with a seq2seq generator.

Coscientist: LLM-orchestrated autonomous chemical synthesis — web search, code execution, and robotic control.

Chemistry agent with 18 expert tools — demonstrates tool use improving accuracy and reducing hallucination.

Framework for building RAG pipelines, tool-using agents, and multi-step LLM workflows.

Data framework for LLM applications: document indexing, retrieval, query engines, and structured data extraction.

Open-source chemistry agent augmenting LLMs with 18 chemistry tools.

Facebook AI similarity search library — the standard for fast approximate nearest-neighbour search over embedding vectors in RAG pipelines.

Benchmark suite of 2700+ chemistry questions spanning safety, reactions, and property prediction — used to evaluate LLM chemistry competence.

Eight chemistry QA tasks for evaluating LLMs: name prediction, yield prediction, reaction prediction, and reagent selection.

1400 scientific claim–abstract pairs with sentence-level rationale labels. Useful for building and evaluating RAG pipelines that retrieve scientific evidence.