Verified Expert in Engineering
Generative Pre-trained Transformers (GPT) Developer
Gabriel completed his master's before spending six years at Amazon working with the Search and Discovery team, and the Music organization. He has nine years of experience building distributed systems with AWS, including natural language processing, speech processing, recommendation systems, and deep learning. Gabriel is a team player who onboarded and mentored several new hires, and junior engineers at Amazon.
Amazon SageMaker, Spark, Jupyter, PyCharm, Git, Linux
The most amazing...
...algorithm I've developed helped millions of Amazon.com customers find the right products out of a catalog of billions of items.
Senior Machine Learning Engineer
GPSE Consulting, LLC
- Created a five-day machine learning engineering workshop to help companies bootstrap machine learning teams internally.
- Customized an online lab platform based on Jupyter to host workshops, including one-touch sign-in with YubiKeys.
- Provided consulting services to companies looking for machine learning expertise.
Machine Learning Engineer II
Amazon Music (A2Z)
- Developed a generative model for automated playlist creation.
- Designed a multi-modal approach to music classification to improve customer experience across several applications.
- Fixed genre misclassification issues for over 100,000 tracks.
- Acted as a machine learning consultant for several external teams.
Software Development Engineer II
- Developed, and maintained a faceted search system with an attributed yearly sale of over $500 million.
- Developed, and deployed a new algorithm for faceted search across several categories and countries.
- Delivered over 10 tech talks on various topics such as AWS, machine learning, and crowdsourcing.
- Mentored five engineers, and interviewed over 130 candidates.
- Organized a Hackday for three consecutive years for an organization of over 300 engineers.
Carnegie Mellon University
- Maintained the Let's Go Bus information system.
- Transcribed over one million speech utterances with crowdsourcing.
- Organized monthly lunches on crowdsourcing for language technologies.
Fixing the Amazon.com Product Cataloghttps://patents.google.com/patent/US20150378975A1/en
I designed and coded several components of the larger pipeline my team was powering. More importantly, I developed a new algorithm for phrase extraction, and scaled it to handle thousands of categories, and billions of products across several marketplaces. EMR (Hadoop, Spark), S3, and several other AWS offerings were integral to the overall architecture, and are now essential elements of my design toolkit.
Knowing that my work had a significant positive impact on several downstream systems (recommendation, search, browse) was a great motivation, but the technical challenge was a driver too. Reprocessing a very large collection of text documents daily, and cost-effectively is not a trivial task. Applying machine-learning at that scale also comes with fun challenges.
Scikit-learn, TensorFlow, Pandas, NumPy
Amazon Elastic MapReduce (EMR), Amazon Elastic Container Service (Amazon ECS), Amazon SageMaker, Git, PyCharm, Jupyter
Data Science, Scrum
Android, Linux, Amazon Web Services (AWS), Amazon EC2, Docker
Amazon S3 (AWS S3)
Machine Learning, Natural Language Processing (NLP), Data Mining, Information Extraction, System Architecture, Big Data, GPT, Generative Pre-trained Transformers (GPT), Deep Learning, Distributed Systems, Recommendation Systems, Cython
Master's Degree in Language Technologies
Carnegie Mellon University - Pittsburgh, Pennsylvania, USA
Bachelor of Engineering Degree in Computer Engineering
École Polytechnique de Montréal - Montreal, Quebec, Canada