Quim Torras
Staff Software Engineer at Google
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolQuim Torras is a Staff Software Engineer based in Zurich with 13 years of experience building production-grade machine learning and audio systems, now leading projects at Google. His background blends deep research in music technology and acoustics (M.A. from McGill) with hands-on engineering—from developing WaveNet implementations and fixing quantization bugs to deploying multi-corpus speech recognition with Kaldi. He has a strong track record in research-engineering roles across academia and industry, contributing to audio analysis and real-time visualization tools, and he brings an uncommon combination of musical sonology and systems modeling to ML problems. Known for improving code readability and model robustness, he pairs rigorous academic thinking with pragmatic, production-focused craftsmanship.
13 years of coding experience
12 years of employment as a software developer
Bachelor of Arts (B.A.), Music (Sonology), Bachelor of Arts (B.A.), Music (Sonology) at ESMUCMaster of Science (M.Sc.), Modeling for Science and Engineering, Master of Science (M.Sc.), Modeling for Science and Engineering at Universitat Autònoma de BarcelonaBachelor of Science (BSc), Audiovisual Systems Engineering, 93%, Bachelor of Science (BSc), Audiovisual Systems Engineering, 93% at Universitat Pompeu FabraMaster of Arts (M.A.), Music Technology, GPA 4.0, Master of Arts (M.A.), Music Technology, GPA 4.0 at McGill University
Catalan, English, Spanish, French
Github Skills (6)
machine-learning10
tensorflow10
python10
model-optimization9
audio-processing9
deep-learning9
Programming languages (9)
Github contributions (5)
ibab/tensorflow-wavenet
Sep 2016 - May 2017
A TensorFlow implementation of DeepMind's WaveNet paper
Role in this project:
ML Engineer
ML EngineerContributions:42 commits, 35 PRs, 38 pushes in 8 months
Contributions summary:Quim primarily focused on improving the readability and structure of the WaveNet model, a TensorFlow implementation. They introduced code style improvements, including enforcing character limits and standardizing comments. Key contributions involve enhancing the `WaveNet` class, adding parameters, and modifying the loss function. Additionally, the user adapted the `generate.py` script, fixing a bug related to quantization steps and improving the output.
deep-learningdeepmindwavenettensorflow
lemonzi/matlab
Mar 2014 - Jan 2018
Contributions:24 commits, 7 pushes, 4 comments in 3 years 11 months
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.