In episode 89 of The Gradient Podcast, Daniel Bashir speaks to Shreya Shankar.Shreya is a computer scientist pursuing her PhD in databases at UC Berkeley. Her research interest is in building end-to-end systems for people to develop production-grade machine learning applications. She was previously the first ML engineer at Viaduct, did research at Google Brain, and software engineering at Facebook. She graduated from Stanford with a B.S. and M.S. in computer science with concentrations in systems and artificial intelligence. At Stanford, helped run SHE++, an organization that helps empower underrepresented minorities in technology.Have suggestions for future podcast guests (or other feedback)? Let us know here or reach us at [email protected] to The Gradient Podcast:  Apple Podcasts  | Spotify | Pocket Casts | RSSFollow The Gradient on TwitterOutline:*

Intro*

Shreya’s background and journey into ML / MLOps*

ML advances in 2013-2016*

Shift in Stanford undergrad class ecosystems, accessibility of deep learning research*

Why Shreya left her job as an ML engineer*

How Shreya became interested in databases, data quality in ML*

Daniel complains about things*

What makes ML engineering uniquely difficult*

Being a “historian of the craft” of ML engineering*

Levels of abstraction, what ML engineers do/don’t have to think about*

Observability for Production ML Pipelines*

Metrics for real-time ML systems*

Proposed solutions*

Moving Fast with Broken Data*

Existing data validation measures and where they fall short*

Partition summarization for data validation*

Small data and quantitative statistics for data cleaning*

Streaming ML Evaluation*

What makes a metric actionable*

Differences in streaming ML vs. batch ML*

Delayed and incomplete labels*

Operationalizing Machine Learning*

The difficult life of an ML engineer*

Best practices, tools, pain points*

Pitfalls in current MLOps tools*

LLMOps / FMOps*

Thoughts on ML Engineering, MLE through the lens of data engineering*

Building products, user expectations for AI products*