Category: Architecture
-
Building Fault-Tolerant AI Pipelines: Smarter Load Balancing and Workflow Orchestration
As AI systems become integral to real-time applications, engineering leaders face the challenge of keeping complex AI pipelines both responsive and resilient. In particular, orchestrating workflows that involve multiple AI models — often calling out to external large language model (LLM) APIs — requires careful design to ensure fault tolerance. This article explores the hurdles…
-
Zero to Hero with Kafka on Kubernetes: Open Source North conf 2021
TLDR; this is a summary of a tech talk I did at Open Source North in 2021. It draws on a past project in which I created an event-driven architecture and patient portal to compare prescription prices by zip code. In the process I figured out an easier way to deploy and manage Kafka on…