10 threads

Ideas followed across the years

Follow a subject through years of MLOps Community sessions: how people approached it, what they tried and where they disagreed. New sessions join each thread as the archive grows.

The archive by year

Session counts and the most common topic tags for each year.

Watching models in production

How do you tell whether a model still works once people depend on it? Speakers look for answers in drift statistics, production traces and evals, including the disputed practice of asking one model to judge another.

352 sessions · since 2020
monitoringobservabilityevals
Feature stores, and second thoughts

Feature stores promised to fill a gap in the ML platform.

75 sessions · since 2020
feature-storesfeature-engineering
The platform team

Who builds the platform data scientists use, and how big should that team be? Then comes the recurring discovery that building it does not make anyone use it.

161 sessions · since 2020
platform-teamsdeveloper-experience
Paying for it

Cost becomes a much more common topic in the archive in 2023.

65 sessions · since 2021
cost
Guardrails, governance and security

What data can a model use, and what should it be allowed to do? Privacy and governance discussions meet newer problems with prompt injection and agent permissions.

218 sessions · since 2020
guardrailssecuritygovernanceprivacy
Humans in the loop

People label the training data and review LLM output.

82 sessions · since 2020
human-in-the-loop
Serving and deployment

The mechanics of getting a model to answer requests, from Flask apps and Kubernetes to inference servers and hosted model APIs.

204 sessions · since 2020
model-servingdeployment
Data quality

The oldest complaint in the archive, restated for every new kind of model.

128 sessions · since 2020
data-qualitydata-pipelines
Agents in production

What can an agent do on its own, and which tools should it get? Once it starts acting, someone needs a way to find out when it goes wrong.

237 sessions · since 2021
agentstool-usemcpcontext-engineering
Open weights against the API

When is a fine-tuned open model a better fit than a hosted API?

110 sessions · since 2020
open-sourcefine-tuning