Person/Lance Martin

• Lance Martin

  • Overview

    • Software Engineer at LangChain focusing on open-source and applied projects
    • Expert in context engineering for AI agents and retrieval-augmented generation (RAG)
    • Created LangChain's first comprehensive course, LangChain Academy, with over 35 lessons on building agents
    • Author of "RAG From Scratch" course featured on FreeCodeCamp with over 450,000 views
    • PhD from Stanford, former tech lead and manager on vision systems in self-driving industry

  • Contact Information

    • GitHub: @rlancemartin
    • LinkedIn: LangChain

  • Popular Repositories

    • auto-evaluator (1.1k stars)

      • Evaluation tool for LLM QA chains
      • Repository: auto-evaluator

    • lex-gpt (343 stars)

      • Language model related repository
      • Repository: lex-gpt

    • generative_agents (96 stars)

      • Forked from joonspk-research/generative_agents
      • Interactive simulacra of human behavior
      • Repository: generative_agents

    • besties-gpt (62 stars)

      • ChatGPT for question-answering on the ALL IN podcast
      • Repository: besties-gpt

    • karpathy-gpt (37 stars)

      • Language model related repository
      • Repository: karpathy-gpt

  • Expertise Areas

    • Context engineering for AI agents
    • Retrieval-Augmented Generation (RAG)
    • Agent development and prompt engineering
    • LLM evaluation and QA chains
    • Computer vision systems
    • Self-driving technology
    • Educational content creation

  • Notable Contributions

    • Developed numerous LangChain Quickstarts on RAG, agents, and prompt engineering
    • Created LangChain Academy with 35+ lessons on building agents
    • Authored "RAG From Scratch" course with 450,000+ views on FreeCodeCamp
    • Uses llms.txt with simple search tools instead of vector store indexing for code agents
      • • This is similar to what Lance Martin talked about on a recent Latent Space/Pod:

          • “What worked really well for me was just passing a simple markdown file—basically a list of docs with URLs and short descriptions—to the code agent. Then it can just grab the right doc and read it when needed. I actually use this all the time, and I don’t even bother with vector store indexing anymore. My go-to is llms.txt with a simple search tool and Claude/Code.” -- #Quote from Person/Lance Martin, LangChain/Team, Context Engineering for Agents on latent.space podcast
            

  • External Links

    • GitHub Profile
    • LangChain Academy Live
    • LangChain LinkedIn
    • Context Engineering for Agents - Latent Space Podcast