CI/CD Fixes Itself Now!- Issue 14
> _ -
Welcome back to VibeOps 👋
This week’s AI 4 DevOps news are so interesting: Pulumi announced an AI platform engineer, AWS is bringing interactive AI into GitHub PRs, and Microsoft is rethinking Logic Apps for agents. Oh, and we found a research framework that uses LLMs to fix CI/CD failures.
Coffee ready? Here are this week's AI 4 DevOps developments.
Pulumi Announced Neo, an AI Platform Engineer
Pulumi announced Neo as an “AI-powered platform engineer” that they claim can accelerate infrastructure automation by up to 18×, cut provisioning times from days to hours, and reduce policy violations with built-in governance.
According to Pulumi, Neo is designed to handle natural language requests (e.g., “upgrade all Kubernetes clusters”), respect compliance policies, understand dependencies, and provide audit trails, effectively acting as an intelligent assistant for managing the full cloud infrastructure lifecycle.
Amazon Q Developer adds a new “interactive code review” in GitHub PRs.
AWS introduced an interactive code review feature in GitHub using Amazon Q Developer.
The feature lets developers request explanations, get suggested fixes, and apply changes directly in pull requests.
🔗 aws.com
Microsoft Previews Logic Apps MCP Servers
Microsoft has launched a public preview of Logic Apps MCP Servers, allowing connectors in Azure Logic Apps to act as reusable MCP tools.
This is supposed to make it easier for AI agents to securely access enterprise systems, automate workflows, and integrate into DevOps processes without extensive custom code.
Upcoming Events
Stakpak Ship It
Missed our last Stakpak Ship It? You missed out on us:
Designed an AWS Infrastructure using stakpak
Optimized an existing AWS Infrastructure using the AWS will architected framework
But don’t worry, our next Stakpak Ship It session is around the corner…
🗓️ Every Wednesday at 10:00 GMT-7 on Discord Live.
P.S.: Don’t forget your coffee
Learning Resources
Every week, we come across interesting blogs, papers, and tutorials that help us level up as a DevOps team. Here are some of the coolest ones we found this week. We hope they inspire you, too.
A Two-Staged LLM-Based Framework for CI/CD Failure Detection and Remediation with Industrial Validation
The authors present LogSage, an end-to-end framework that uses large language models (LLMs) to help detect when CI/CD pipelines fail, figure out why, and suggest / generate fixes automatically. It works in two stages: first, it preprocesses logs to extract the relevant error segments and remove noise, then it uses retrieval-augmented generation (RAG) plus tool-calling (i.e. invoking automated fix routines) to propose solutions based on past resolution patterns
A quick favor (helps a lot):
To make sure this doesn’t land in spam…
Reply to this email and just say: “Got it” or “Let’s gooo” or drop your favorite emoji.
That tiny reply tells your inbox you actually want this — and it tells me you’re real. btw, we read every single message.
One Last Thing
VibeOps is made by the community for the community, so reply to these emails with all your cool AI use cases for DevOps and any cool AI tools that you use, and we will feature them in our next issue
Written with 💚 (and a lot of synthetic caffeine) by the VibeOps team ☕🤖