Suggestions
Maria Khalusova
Developer Advocate at JetBrains
Maria Khalusova is a technology professional with a diverse background in software development and developer advocacy. She is currently a Senior Developer Advocate at Iterative.ai, where she focuses on promoting best practices from traditional software development, such as version control and continuous integration.1
Prior to her current role, Maria held the position of Developer Advocate at JetBrains, where she specialized in data science, data engineering, and machine learning.36 Her work at JetBrains involved making machine learning more accessible on the Java Virtual Machine (JVM) using Kotlin.6
Maria's career has included various roles such as Developer Advocate, Product Manager, Product Marketing Manager, and Technical Writer. She has a keen interest in emerging technologies, particularly in the field of artificial intelligence and machine learning.
Recent Activities
Maria has been actively sharing her knowledge and experiences in the tech community:
- In August 2024, she shared a nostalgic post about her first tech blog post, which focused on the code coverage feature in IntelliJ IDEA 6.0.2
- In September 2024, she wrote a blog post about embedding models for RAG (Retrieval-Augmented Generation) and shared a notebook on comparing embedding models.4
Current Focus
As of the most recent information available, Maria is working at Unstructured.IO and describes herself as an "LLM whisperer" and "RAG tinkerer".5 This suggests her current work involves large language models and retrieval-augmented generation techniques.
Maria Khalusova continues to be an active voice in the developer community, sharing insights and knowledge about various aspects of software development and machine learning technologies.
Highlights
Our new website looks sooooo good! 😻
Asking “What is the best chunk size for RAG?” without any additional context is like asking, “What’s the best thing to wear?” Wear where? What’s the weather like? What size are you? Are you going to a wedding or hiking a trail? There’s no single answer that works for every situation.
Same goes for chunking. The “best” chunk size depends on a range of factors, and without those, the question is incomplete. Here are some of the questions to ask instead:
- What does your data look like? Financial statements, technical manuals, customer support transcripts are not the same. They all vary in structure, style, and length.
- What is your use case? Are you trying to answer questions with specific facts? Are you gathering multiple documents to summarize for a report? Do you pull from transcripts and need to preserve speaker attribution?
- How much context do you typically need to retrieve to satisfy a typical query? Simple facts may only require a sentence or two. Creative tasks may require larger context. Analytical queries may need a whole bunch of supporting evidence.
Once you have some answers to these, you can get further into the technical weeds and experiment with chunking to find an optimal size. Bottom line, however, is - there's no universal "best" chunk size. RAG exists to solve different problems across varied domains. Understand the problem you’re solving and look at your data.
