LangChain, a leading platform in AI development, has released its latest update, introducing new use cases and improvements across the ecosystem. According to the LangChain blog, the update includes advancements to LangGraph Cloud, LangSmith’s self-improving evaluator, and revamped documentation for LangGraph.
LangSmith: Self-improvement evaluator
LangSmith has introduced a significant improvement that allows humans to correct the “LLM-as-Judge” assessment. This feedback loop is designed to improve the accuracy of future assessments by incorporating human corrections in a few shot examples. Users can refer to the demo video for integrating the self-improving evaluator into their dataset.
LangGraph Cloud: Diverse Use Cases
LangGraph Cloud continues to expand its utilities for running large-scale LLM applications. Notable use cases include building a full-stack generational UI app, deploying a Discord bot that learns from conversations, and creating a self-correcting RAG application to effectively handle model hallucinations. Detailed guidance and examples can be found in our various video tutorials.
Revised LangGraph documentation
The LangGraph documentation has been revamped to provide clearer and more actionable guidance. The new sections include:
Upcoming Events and Hackathons
LangChain will host an Agents Hackathon on August 11th in San Francisco, featuring talks from industry leaders and opportunities to win prizes and credits. The event aims to foster innovation and collaboration among AI developers. Interested participants can apply here.
In case you missed it, LangChain recently hosted regional meetups in NYC and Austin, Texas, bringing together builders and enthusiasts. The panel discussion featuring Edo Liberty (CEO of Pinecone) and Harrison Chase (CEO of LangChain) can be rewatched here.
Customer Success Stories
LangSmith was adopted by Wordsmith, the legal team’s AI assistant, to optimize the product life cycle from debugging to production. The platform enabled Wordsmith to set test criteria and achieve rapid iterations, resulting in higher accuracy and recall rates. Read the full story here.
New Computer, the creator of the personal AI assistant Dot, used LangSmith to enhance the agent memory system, resulting in significant performance improvements. You can read more insight into their approach here.
For more updates and detailed guides, users are encouraged to visit the LangChain blog and the official LangChain YouTube channel.
Image source: Shutterstock