Richhiey Thomas is a pattern recognition and data engineer specializing in music-focused AI, currently driving MyGroove research at Red Bull with 11 years of software and data experience. He holds an M.Sc. in Data and Knowledge Engineering from Otto-von-Guericke University and built his expertise developing music generation and transcription models during roles at OVGU AILab, Skoove, and PhonicScore. His background blends research and product: from a thesis on compressed transformers for piano transcription to production data engineering and data science work at consumer-facing music tech companies. An active contributor to open-source search infrastructure, he extended Xapian with regex matching, clustering APIs, and performance improvements—demonstrating a knack for scalable back-end systems. Based in Mumbai with international academic experience in Germany, he pairs deep learning research with practical engineering to move creative AI from prototype to product. Notably, he’s uniquely comfortable navigating both low-level C++ search engine internals and high-level sequence models for music.
11 years of coding experience
7 years of employment as a software developer
M.Sc. in Data and Knowledge Engineering, M.Sc. in Data and Knowledge Engineering at Otto-von-Guericke University Magdeburg
Junior College, Sciences, Junior College, Sciences at St. Xavier's College
School, School at Don Bosco High School, Matunga, Mumbai
Bachelor in Computer Engineering, Bachelor in Computer Engineering at Mumbai University
Mirror of the Xapian repository. You're welcome to open pull requests on github (they'll just get merged indirectly).
Role in this project:
Back-end Developer
Contributions:11 commits, 20 PRs, 193 comments in 1 year 4 months
Contributions summary:Richhiey primarily focused on extending the functionality of the Xapian search engine, implementing features for omegascript, the query language. They added a `$match` command for regex matching, expanding the capabilities of omegascript. Furthermore, the user optimized the codebase by replacing `std::set` with `std::unordered_set` in the `SimpleStopper` class, and also added and modified the clustering API. The user also made changes to handle stopword removal.
Training FolkRNN models to generate Irish double jigs for the AI Music Generation Challenge 2020 as part of the Joint Conference on AI Music Creativity (CSMC + MuMe), KTH Stockholm
Contributions:79 commits, 12 PRs, 90 pushes in 2 years 2 months
mumeabc-notationmusic-generationstockholmirish
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.