Why Extend Went All-In on Serverless Platform Engineering
MAR 4, 202662 MIN
Why Extend Went All-In on Serverless Platform Engineering
MAR 4, 202662 MIN
Description
Billions of requests a month on AWS Lambda can cost less than a single engineer’s laptop budget, but only if the architecture and developer workflow are designed for it.Justin Masse, Senior Platform DevOps Engineer at Extend, shares how Extend committed early to a serverless-first approach and built a platform that prioritizes developer speed and low operational toil. The conversation breaks down what it takes to run active-active, multi-region systems in a serverless world, how the team keeps services small and fast, and why asynchronous, event-driven design changes both reliability and cost.You’ll also hear how Extend treats developer experience as a core platform responsibility: templated microservices, fast deployment pipelines, ephemeral environments for pull requests, and infrastructure that developers can own without becoming cloud specialists. A big theme is using AWS CDK and internal abstractions to keep infrastructure close to the application code, so teams can move quickly while keeping platform standards consistent.Finally, the discussion gets practical about tradeoffs that show up after the “serverless is easy” pitch: local development challenges, the real cost center (observability), and where AI is helping today, including an internal agent that diagnoses failed deployments and suggests fixes.What you’ll learnWhy Extend avoids servers and VPC complexity, and what they use insteadPatterns for active-active, multi-region thinking in a serverless architectureHow DevEx practices like templates and ephemeral environments reduce frictionA pragmatic approach to IaC with CDK and reusable internal constructsWhere serverless costs stay low, and why observability often dominates the billHow AI is being applied to platform workflows without skipping engineering judgmentGuest: Jusin Masse, Senior Platform DevOps Engineer at ExtendJustin Masse is a self-proclaimed lead chaos engineer, recognized within niche engineering communities for his expertise Chaos Engineering and Infrastructure & DevOps.The father of three young kids, a husband, a recent MBA graduate, recent cancer survivor, and competitive powerlifter, he still finds time to actively contribute to the platform engineering community.Justin Masse, websiteJustin Masse, GitHubExtend, websiteLinks to interesting things from this episode:Episode with Adrian Cockroft“From $erverless to Elixir” by Cory O’Daniel