I should start by saying theyorubayesian
is a portmanteau of Yorùbá and Bayesian.
As theyorubayesian, I am a software engineer operating/researching at the intersection of distributed systems, information retrieval, and artificial intelligence.
You may learn about other parts of me in time.
See Google Scholar for research publications.
I started my career building and managing machine learning systems in on-premise and hybrid cloud environments for enterprises. I have expertise in helping companies evaluate their machine-learning technical debt and production readiness. I also help them achieve continuous integration, training, deployment, and monitoring goals. In 2020, I was a technical mentor for Udacity, guiding professionals taking their Azure Machine Learning Engineer NanoDegrees, providing feedback on projects and helping them build an intuition of Azure's Machine Learning platform. This role deepened my understanding of the challenges professionals and companies encounter when using machine learning in production.
My long-term goal is to advance technology that exists for African languages. I believe in the transformative power of this goal for the continent. I have used my web crawling experience to help develop test collections for text classification, cross-lingual information retrieval and to scale the pre-training data available for 16 African languages. With The African Research Collective (TARESCO), I am working on projects whose successes are critical for state-of-the-art technology for these languages to exist, including but not limited to language identification, speech transcription, optical character recognition to obtain data, and more. Feel free to reach out to sponsor TARESCO or collaborate in this area :)
I was the Lead Instructor for the 2019 AI Saturdays Lagos' Cohort. I worked with instructors, mentors and buddies to develop a hands-on introduction to data science course for over 100 students. I still interact with students who got their start through AI Saturdays Lagos and provide guidance where possible.
I also led a Neural Nets for Natural Language Processing Reading Cohort with the Masakhane community. I hope to champion more reading cohorts with an improved format in future. I developed an AutoML course for LinkedIn Learning. While that was fun and challenging, I prefer interaction-based learning, so it may be my solo effort in that regard.