About Me

Hi! I’m Diego, an engineering leader focused on information retrieval, ML ranking, and retrieval-augmented LLM systems.

I’m currently an AI Engineering Manager at Bloomberg, working on production LLM applications, agentic search, and retrieval systems.

Before moving into management, I spent several years as a software engineer at Bloomberg working on search & ranking — including designing and open-sourcing Apache Solr’s Learning to Rank module, which was adopted into the open-source core.

In a previous life I was a Post-Doctoral Fellow at the Institute for Advanced Studies of Lucca (IMT) and an Associate Researcher at the High Performance Computing Laboratory (HPC), National Research Council (CNR).

Disclaimer: The views and opinions expressed on this website are my own and do not necessarily reflect the views or opinions of my employer.

Contacts / Social stuff

• Email: diego.ceccarelli@gmail.com
• LinkedIn
• GitHub
• Google Scholar
• My favorite songs on Spotify

Education

• Ph.D. in Computer Science — Information Retrieval, University of Pisa, 2013. 20 peer-reviewed publications. (Full pdf of the thesis - Abstract)
• M.Sc. cum laude in Computer Science, University of Pisa, 2009.
• B.Sc. cum laude in Computer Science, University of Pisa, 2006.
• Bachelor of Music in Flute Performance, Musical Institute Pietro Mascagni, 2006.

Publications

You can find my publications on Google Scholar.

Datasets

• SAC 2013: When Entities Meet Query Recommender Systems: Semantic Search Shortcuts dataset

(free) Software

I enjoy contributing to open source in my spare time.

In 2016, I contributed to the release of the Learning to Rank module for Apache Solr. I also gave several talks on how learning to rank works, how to use the Apache Solr module, and the challenges of releasing learning to rank in production.

When I’m in the mood I maintain json-wikipedia, a library to convert Wikipedia XML dumps into a JSON/Avro semi-structured format.