Uploaded October 2024 | Updated November 2026, 2 hours ago
One of the benefits of working in big tech is you often build or see toolings and solutions internally that could benefit a broader audience. Whether it's a query engine or a data storage format.
Richard Meng has been working as a Staff Software Engineer at companies like Linkedin and Snowflake with data as a particular focus.
Most recently his experiences at Snowflake provided him with several insights into how LLMs can better be used to process data. In particular, unstructured data.
For example, instead of going the standard method of chunking the documents and embedding them into vectors to allow future users to ask questions of said data. Richard's team believes in directly using vision LLM on unstructured data in their raw format, which preserves the original context of said unstructured data as much as possible.
In this live chat, I'll be focusing on Richard's experiences that led up to this point and how he is viewing LLMs impacting the work of data engineers.
If you'd like to reach out to Richard and ask some questions, you can reach out here:
linkedin.com/in/berkeleymeng
Also, if you are looking to analyze PDFs or Images with SQL, you can try out Roe here.
getroe.ai
Disclosure: Seattle Data Guy does have a stake in Roe.AI
One of the benefits of working in big tech is you often build or see toolings and solutions internally that could benefit a broader audience. Whether it's a query engine or a data storage format.
Richard Meng has been working as a Staff Software Engineer at companies like Linkedin and Snowflake with data as a particular focus.
Most recently his experiences at Snowflake provided him with several insights into how LLMs can better be used to process data. In particular, unstructured data.
For example, instead of going the standard method of chunking the documents and embedding them into vectors to allow future users to ask questions of said data. Richard's team believes in directly using vision LLM on unstructured data in their raw format, which preserves the original context of said unstructured data as much as possible.
In this live chat, I'll be focusing on Richard's experiences that led up to this point and how he is viewing LLMs impacting the work of data engineers.
If you'd like to reach out to Richard and ask some questions, you can reach out here:
linkedin.com/in/berkeleymeng
Also, if you are looking to analyze PDFs or Images with SQL, you can try out Roe here.
getroe.ai
Disclosure: Seattle Data Guy does have a stake in Roe.AI










