MLOps Community

The MLOps community is an open and transparent community where all are welcome to participate. It is a place where MLOps practitioners can collaborate and share

[MLOps Community @mlopscommunity]

Coding Agents Conference. March 3rd at the Computer History Museum. https://t.co/i1HyBy4Agp

[MLOps Community @mlopscommunity]

One comment that will probably spark debate: "Complexity should never be planned. It should be a result of time."

https://t.co/L5O1mafEq1

Do you buy that approach, or do some agent systems need more structure from day one?

[MLOps Community @mlopscommunity]

Spent some time listening to Thiago Cardoso, Director of Data & AI at @iFood, talk about what happens when agents move beyond demos and start handling money, payroll, credit, and real business operations.

[MLOps Community @mlopscommunity]

3⃣ His take on agent architecture was surprisingly simple: start with one agent. Add complexity only when evals show something is breaking. The structure emerges from failures and feedback, not from designing a giant multi-agent system upfront.

[MLOps Community @mlopscommunity]

Registration: https://t.co/xZKTpq4UjS

[MLOps Community @mlopscommunity]

Most enterprise RAG systems have a dirty secret.

They retrieve based on relevance, not authorization.

That means the document most likely to answer your question might also be the one your agent should never see.

[MLOps Community @mlopscommunity]

✅What "secure-by-default" agent architectures might look like

🗓️ June 11
🕜1:30 PM ET

Join us if you're building agents, RAG systems, internal copilots, or simply wondering whether today's AI architectures are ready for enterprise reality.

[MLOps Community @mlopscommunity]

🎟️ Register here:
https://t.co/4Eg33LjC6w

#MLOps #MachineLearning #FeatureEngineering #DataEngineering #RealtimeML #Chronon #Zipline #AIEngineering

[MLOps Community @mlopscommunity]

⚡ Building Real-Time ML Systems with Zipline + Chronon ⚡

Real-time ML sounds great in theory. In practice, it's a constant battle against feature freshness, low-latency serving, training-serving skew, and increasingly complex data pipelines.

[MLOps Community @mlopscommunity]

📅 June 10
🕤 9:30–10:30 AM PT
💻 Virtual

Whether you're building recommendation systems, fraud detection pipelines, or next-generation AI applications, this conversation will provide practical insights from teams operating real-time ML at scale.

[MLOps Community @mlopscommunity]

@hvent90 @shcallaway @sazabi I'll take it.

[MLOps Community @mlopscommunity]

Had @shcallaway on the podcast to talk about @sazabi, the observability platform he's building, and he's openly gunning for Datadog. Most of the hour was him defending three positions that cut against how teams have done observability for the last decade.