Code Remix Summit’s Post

Dinesh A., Principal Engineer at Wells Fargo, is breaking down feature flags at Code Remix Summit. 🏳️ Not just what they are, but how to actually use them without creating chaos. In his session, "Feature Flags: How to Ship Faster, Safer, and Smarter," get best practices that prevent flag sprawl and technical debt. Hear real case studies showing how companies are using flags to ship with confidence. 💪 Grab your ticket ➡️ https://lnkd.in/g5r2MUSJ #CodeRemixSummit #FeatureFlags #WellsFargo #ContinuousDelivery #DevOps

Random Labs Unveils Slate: Open-Source CLI Agent for Large-Scale Software Engineering 📌 Random Labs launches Slate, an open-source CLI agent built for massive codebases - handling up to 2 million lines of code with persistent context and infrastructure integration. Unlike chat-based AI, Slate automates debugging, testing, and deployment via terminal workflows, slashing latency for enterprise-scale engineering. It’s a game-changer for DevOps teams needing deep, long-running analysis without GUIs or complex setups. 🔗 Read more: https://lnkd.in/dR-WKw7Y #Randomlabs #Slateagent #Cliengine #Codebasemanagement #Largescaledev

Your coding agent ships features in days. But hand it a P1 at 3 AM and it struggles. Why? Production engineering is a completely different game: - 10x more services - 10x more tools - 10x more people involved - 10x more time urgency - No public data to train on Those multipliers compound into a 1000x search space. Coding is single-player. Production is a multiplayer war room. Utkarsh O. wrote about why and what it actually takes to close the gap. https://lnkd.in/g8md4nxh #AISRE #IncidentManagement #DevOps #SRE #ProductionEngineering #Autoheal #AICodingAgents

Ever since Opus 4.5 came in November, I have been a Claude Code fanboy. Almost a bit too much. Yet, I can confidently say that solving the Production Engineering problem is a different ballgame than solving the coding problem. Think Context, Security and Collaboration. Utkarsh takes you through a ride on why so - MUST READ!

Most of the time when we talk about security, it’s treated like a final checklist item before deployment. After watching a deep dive on DevSecOps today, I realized how flawed that approach actually is. Security isn’t something you “add later”. It has to be part of the entire pipeline. The flow made a lot of sense to me: Code → CI Check → Build → Test & Scan → Deploy And at every step, there’s some form of security involved: • Code level checks • Dependency scanning • Image scanning (like using Trivy) • Secure deployments What I liked was that even a simple app can be set up with a proper CI/CD pipeline + security checks. Coming from a backend + fintech background, this felt very relevant. Because in payments, even a small issue can create serious problems — not just bugs, but actual financial impact. One thing that stuck with me: If security is not automated, it will be skipped. Planning to explore this more, especially around: • Secure CI/CD pipelines • Docker + image scanning • Kubernetes deployments Still learning this space, but it’s definitely something I want to get better at. How are you guys handling security in your workflows? #DevSecOps #BackendEngineering #SystemDesign #FinTech #CI_CD #Docker #Kubernetes

One of the hardest problems in backend engineering is not writing code. It is keeping architecture consistent as teams scale. Coding agents help by enforcing Domain-Driven Design automatically: • Understanding bounded contexts • Generating code only within allowed domains • Preventing cross-domain leakage • Enforcing folder structures and naming conventions • Maintaining repository → service → controller layering • Generating domain-aligned tests Agents are no longer just code assistants. They become architecture enforcers by default. Build once. Enforce always. 🌐 www.kadellabs.com 📩 contact@kadellabs.com #DomainDrivenDesign #CodingAgents #BackendArchitecture #SoftwareEngineering #ScalableSystems #AIInEngineering #KadelLabs

A lottttt has happened in the software development space the last 2 months. Lots of unknowns and anxiety and new threat vectors. BUT, the data shows that Engineering roles are growing.

The most expensive bug I ever found was a "Heisenbug." It passed every local test. It passed the CI/CD pipeline. It even passed a week of staging. But the second we hit 1,000 concurrent users in production? Total gridlock. We were hit by a Race Condition. That is the nightmare scenario where two threads fight over the same piece of memory and everyone loses. If you are still trying to catch these by "looping a test 100 times" or adding Thread.sleep(2000) to your scripts, you are not testing. You are just procrastinating. Here is how we actually hunt them down now: • Stop Being "Nice" to Your Code: In automation, we often create "perfect" environments. In the real world, the network jitters and CPUs throttle. I started using tools like Gremlin to purposely slow down specific microservices. If your "Service A" assumes "Service B" will always be fast, chaos engineering will expose that lie in minutes. • The "Sharded" Stress Test: Instead of running tests one by one, we now fire off 50 or 100 instances of the exact same test simultaneously against a shared database. If there is a row locking issue or a transaction isolation failure, this brute force approach drags it into the light. • Trust the Auto Wait: Modern tools like Playwright are great because they do not use fixed timers. If a test is flaky even with auto waiting, do not just retry it. That flakiness is usually a signal that your frontend and backend are not syncing correctly. The Lesson: If your automation environment is too "clean," it is lying to you. Production is messy, loud, and unpredictable. Your tests should be, too. How do you handle concurrency? Do you use a stress and observe approach, or are you moving toward deterministic simulation? Let’s swap horror stories in the comments. #SoftwareEngineering #Automation #Programming #QA #DevOps #TechLife

With coding becoming more and more easier, devops needs to catch up with the pace of development. However, the world of devops is different and the signals are very different. This is when CodeKarma becomes super relevant. I will be there at the DevOpsConclave. Do drop in for a chat.

CodeKarma is bridging that gap by providing production context right where developers live—in the IDE. No more guessing games, just better automation and observability that actually tells the truth. Check out the CodeKarma team at DevOpsConclave with UBS Forums to see how they're making "shipping to prod" a lot less scary. #observability #devops #reliability #Platformengineering #GSI #vibecoding #SRE #monitoring #developer #tools