Enhancements for FAQ Chatbot for Vitess

programs/summerofcode/2025.md

@@ -62,3 +62,30 @@ You can find the project ideas from previous year [here](./2024.md).
   - Valentin Delaye (@jonesbusy, [email protected]) - primary
   - Feynman Zhou (@FeynmanZhou, [email protected])
 - Upstream Issues: https://github.com/oras-project/oras-java/issues
+
+#### Vitess
+
+#### Enhancements for FAQ Chatbot for Vitess
+
+Vitess is a distributed database system built on MySQL. Developers often need to search through documentation, Slack
+discussions, and GitHub issues to find answers. We are starting a project to implement an AI-powered FAQ chatbot using
+**Retrieval-Augmented Generation**, integrating **vector search** with an **LLM** (such as OpenAI, DeepSeek,
+GPT-4, Mistral, Llama 3). The chatbot will be available via a **CLI and Slack bot** for developer support.
+
+In the next phase, which will be implemented in this Summer Of Code (SOC) project, we will be adding more features like:
+* Content filtering for chatbot safety and response validation
+* Fine-tuning the model for improved accuracy
+* Pipelines for refreshing data from new/updated docs
+* Caching previous replies to reduce LLM costs
+* Rate-limiting
+* Better benchmarking for iterative improvements
+* User feedback integration to improve relevancy
+
+
+- Expected Outcome: Improved chatbot that provides accurate Vitess-related answers via CLI and Slack, using indexed documentation and discussions for retrieval.
+- Recommended Skills: golang, python, LLM APIs, vector databases
+- Expected project size: large (~350 hour projects)
+- Mentor(s):
+  -  Rohit Nayak (@rohit-nayak-ps, [email protected])
+  -  Manan Gupta (@GuptaManan100, [email protected])
+- Upstream Issue: https://github.com/vitessio/vitess/issues/17690