🦸 Community Superhero
Our community superhero this week is Aimee Barciauskas, data engineer at Development Seed. Aimee helps shape how Earth observation data is accessed in the cloud, contributing to platforms like NASA’s VEDA and serving on the Pangeo steering committee to make large-scale environmental datasets more accessible and interoperable. She has also played a role in helping scientists and engineers collaborate on cloud-native approaches to planetary-scale data. If you care about modern cloud computing being used for real-world impact, Aimee is someone worth following. Thank you so much for everything you do, Aimee!
💯 Spotlight
My favorite post from last week is from Ran Isenberg. He shared an honest experience of rebuilding his personal site with Claude while trying to figure out if “SaaS is dead.” While we know SaaS is certainly not dead (and Ran confirms it in his article), the big takeaways from this post are how he approaches the build with managed expectations and a more subtle, inferred thought that engineers can and should build hyper-personalized software for themselves. There will always be a need for SaaS that solves a problem better than anybody else. But for smaller things that make your daily life easier, we’re in a golden era. The site looks beautiful, Ran! Great job!
🔥 My Favorite Content
While we’re talking about SaaS, an important component of any SaaS app is tenant isolation. Lambda released support for tenant isolation at the end of 2025, enabling apps to keep execution environments contained to a single tenant. How, you ask? Pubudu Jayawardana tells us exactly how in his post where he covers everything about Lambda tenant isolation. He covers how this is an improvement over the unscalable workarounds of the past, how to wire it up to API gateway, and some improvements he’d like to see in the future. This is a great article with practical lessons if you’re looking to try it out for real.
2026 is already shaping up as the year we get serious about building with AI. Articles, social posts, and conference talks are turning away from “look at what’s possible” to “here’s how you do this in production.” Along these lines, J Simpson published an article on how to reduce API latency in agentic AI systems. Reading through these, they make sense as mindset shifts to how we approach agent tooling. Some of his points you’ve definitely heard before, like caching is more important than ever, and others you might have heard me talk about on a livestream, but there’s a few novel ideas in there that really elevate how we build with agents.
For more on production-ready agents, Jeremy Daly published some light reading last week about context engineering for commercial agent systems (just kidding, it’s an incredibly long and thorough post that should be a book, hint, hint, Jeremy). I honestly can’t summarize all the pieces in here - you’ll have to check it out for yourself. I can, however, tell you that it reframes agentic mental models for me. The post has a ton of detail, but is mostly about managing context safely at scale, i.e. controlling what the model knows, remembers, costs, and is allowed to do. It’s a big dose of reality coming from lived experience from a seasoned pro. Well done, Jeremy!
I apparently went on a tear last week. I published two blog posts on performance engineering, one on routing layers in large-scale caching systems and another on the NxM problem in distributed systems. Luckily for most of us in this crowd, serverless handles these problems decently well. But when you’re responsible for large fleets of containers and/or pods, these issues pop up quick as you scale. I also released a blog post introducing Valkey admin, a graphical interface for operators to manage their cache clusters. And finally, I was on Serverless Office Hours with Andres Moreno talking about a serverless video content repurposer for content creators we built late last year. We went over the architecture, some fun workarounds we had to do to make it fully serverless, and talked about productionizing AI agents as well.
💡 Tip of the Week
This specific issue of the newsletter feels like reassurance that AI will not be taking our jobs in the near future - probably not at all. But if you still aren’t convinced, here’s Jeff Barr with the same message saying that AI is taking engineers to new heights as long as we keep improving our communication skills, providing context, and thinking creatively.
🐣 New Releases
Reminder, all releases from AWS can be found on AWS News by Luc van Donkersgoed. Below are my favorite from last week.
Lambda Durable Functions is now available for Java functions. This is via the Durable Execution SDK which is in preview. Great to see expanding language support!
AWS launched a playground for Aurora DSQL database exploration without requiring an AWS account. The playground, found here, let’s you start with a blank state or with a couple of functional examples.
An interesting new service, AWS Elemental Inference, was announced - a service that transforms live landscape-oriented video to portrait in real-time. This allows creators to reach their audience everywhere, instantly. I like the idea (I think).
Last Words
Wow things are really picking up! I had to go through a lot of content this week to find this for y’all. I love the spike in motivation and the genuine content being created. Keep up the good work!
What did I miss? What made you nod along (or 🙄)? Hit reply if you’re reading the email. Prefer socials? Ping me on Twitter, LinkedIn, or email.
Happy coding!
Allen