Sign In

Sean Owen

Big-data data science and engineering leader, writer, founder

Sean Owen is a prominent Staff Research Scientist at Mosaic, a division of Databricks, where he specializes in data science and machine learning. Based in Austin, Texas, he has been with Databricks since August 2018, initially serving as a Principal Specialist for Data Science and Machine Learning. His role involves leading architectural advisory for data science and machine learning initiatives and delivering thought leadership in these areas.12

Professional Background

  • Databricks: As a Principal Specialist, Owen focuses on providing technical guidance in data science and machine learning. He also represents customer needs to product and engineering teams.
  • Cloudera: Before joining Databricks, he was the Director of Data Science at Cloudera from 2013 to 2018, where he led global customer-facing data science engagements.
  • Myrrix Ltd: He founded this company to work on large-scale machine learning using Apache technologies, which later became part of Cloudera's Oryx project.
  • Google: Owen co-led mobile web search initiatives and contributed to the development of the Android Barcode Scanner during his tenure from 2004 to 2008.12

Education

  • Harvard University: Bachelor’s degree in Computer Science, graduating magna cum laude (1996-2000).
  • London Business School: MBA with a focus on Entrepreneurship (2008-2010) .1

Contributions and Recognition

Owen is recognized as a committer and Project Management Committee (PMC) member for Apache Spark. He has received accolades such as the 2022 SIGMOD Systems Award for contributions to Apache Spark and has been named a Quora Top Writer multiple times between 2014 and 2017.12

In summary, Sean Owen's extensive experience in data science, machine learning, and software engineering positions him as a key figure in advancing these fields within the enterprise landscape.

Highlights

Feb 1 · linkedin

https://lnkd.in/gPD8dRVU

New LLM just dropped courtesy of the Allen Institute for AI (AI2). The transparency in training code and dataset, and attention to training impact, is refreshing. Oh and the model is good: https://lnkd.in/g3CNiJba

Aug 17 · linkedin

A request for comments on embedding providers!

Embeddings are obviously a vital part of the 'gen AI' wave - they're not just part of classification use cases, but now vector DB retrieval for retrieval-augmented generation with LLMs, and more.

OpenAI (and other providers) offer embeddings as a service. There are also lots of OSS embeddings. While OpenAI LLMs are still just about the state of the art, it looks like OSS embeddings outperform it and other proprietary models. Here's at least one benchmark, MTEB: https://lnkd.in/gw7dHhGC

I've heard concerns about using embeddings-as-a-service, because any change to the third-party embedding renders any data that's been embedded with it useless. Could be a big deal if you've spent time/$ to embed a billion docs. I believe another similar service had to help customers re-embed data after such a change.

So the question is: is there a reason to use OpenAI Embeddings, versus just running e5-large-v2 or instructor-xl or similar yourself? The idea being, seems like they're 'better' and you of course then are not exposed to updates if you don't want to be.

Related Questions

What are Sean Owen's main responsibilities at Databricks?
How did Sean Owen contribute to Apache Spark?
What is the Oryx project that Sean Owen founded?
What is the significance of Sean Owen's role as a Principal Specialist for Data Science and ML at Databricks?
How did Sean Owen's experience at Google influence his career?
Sean Owen
Sean Owen, photo 1
Sean Owen, photo 2
Add to my network

Location

Austin, Texas, United States