Practical MLOps for better models
Abstract
Machine learning operations (MLOps) are often synonymous with large and complex applications, but many MLOps practices help practitioners build better models, regardless of the size. This talk shares best practices for operationalizing a model and practical examples using the open-source MLOps framework vetiver to version, share, deploy, and monitor models.
View the repository || Watch the recording
Post-talk notes
I will always remember PyData Global 2022 as the conference that Hadley Wickham, who is essentially the face of the R language, talked about Python. It was a delightful crossover episode in my world. My talk really was an ode to how I learned about MLOps. It goes through what I thought data science looked like when I learned about it in school, only to be SHOCKED that I needed a little bit of practical MLOps knowledge to collaborate with teammates, not lose models into the abyss of version1, version2, version2_final, version_final_forreal, and overall be able to do my job effectively.