Real World Serverless with theburningmonk

In this episode, I spoke with Matthew Bonig, who is an AWS DevTools Hero and co-author of The CDK Book.We talked about the Dos and Don'ts of CDK, and Matthew shared some tips on testing and how to organize your CDK project.Links from the episode:The CDK BookAdvanced CDK CourseConstruct HubCDK PatternsCDK Community Survey 2023cdk.devWing LangAmptCDK Days track 1CDK Days track 2You can find Matthew on X as @mattbonig.-----For more stories about real-world use of serverless technologies, please subscribe to the channel and follow me on X as @theburningmonk.And if you're hungry for more insights, best practices, and invaluable tips on building serverless apps, make sure to subscribe to our free newsletter and elevate your serverless game! https://theburningmonk.com/subscribeOpening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

Charity Majors, co-founder and CTO of Honeycomb, shares insights on Parse's acquisition by Facebook, observability, devops, deploying on Fridays, and the significance of team safety and celebrating failures for success.

In this episode, I spoke with Gerard Sans, who is a Google Developer Expert and my co-author for the AppSync Masterclass.We talked about Google's new AI platform, what it entails and how Google's AI offerings compare to OpenAI. For example, what can you do with Bard and Duet AI and why one might consider using them instead of ChatGPT or CoPilot. We discussed Midjourney vs. Stable Diffusion, and how platforms need to be more responsible with their AI offerings. And we also touched on the idea of AI agents and once again, LangChain was part of the conversation.You can also watch this episode on YouTube here and see Gerard's slides and demos.Links from the episode:Google's MakerSuite announcementAI for Google WorkspaceStable DiffusionStable Diffusion's new DreamBooth featureAdobe FireflyMidjourneyMidjourney's new inpainting featureHow ElevenLabs's voice cloning made me sound ScottishGoogle BardAnthropic Claude 2Ep81: AI development with LangChain with Matt CareyYou can find Gerard on X as @gerardsans.-----For more stories about real-world use of serverless technologies, please subscribe to the channel and follow me on X as @theburningmonk.And if you're hungry for more insights, best practices, and invaluable tips on building serverless apps, make sure to subscribe to our free newsletter and elevate your serverless game! https://theburningmonk.com/subscribeOpening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

Heitor Lessa, principal solution architect at AWS, talks about the challenges and learnings from building the Lambda Powertools project. They discuss the trade-offs of monolithic Lambda functions and the use of Lambda Layers. Heitor also announces upcoming features and discusses the importance of working in public and building internal communities.

In this episode, I spoke with Matt Carey, who's a senior engineer at Aleios, a premium AWS consultant that focuses on serverless technologies.The focus of our conversation was LangChain, Matt explained what it is and demoed some AI applications that he has developed with LangChain. Such as the popular Quivr app and Code Review GPT. If you want to see the demos in their full glory, then please go to YouTube and watch the episode here.We talked about some common challenges with AI development, such as testing, and how Matt and his team have developed a novel approach for dealing with the non-deterministic nature of these applications.Links from the episode:* Quivr* Code Review GPT* LangChain docs* PrivateGPT* Theodo's OrionTools* Blog on testing Gen AI applications* GenAI days eventsYou can find Matt on X as @mattzcarey and his LinkedIn profile here. You can also learn more about Aleios and their work at https://www.aleios.com. For more stories about real-world use of serverless technologies, please subscribe to the channel and follow me on X as @theburningmonk.And if you're hungry for more insights, best practices, and invaluable tips on building serverless apps, make sure to subscribe to our free newsletter and elevate your serverless game! https://theburningmonk.com/subscribeOpening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

Randall Hunt, VP of Cloud Strategy and Innovation at Caylent, discusses AWS Bedrock and its advantages, debunking AWS's supposed weakness in AI. They explore tokenization, vector stores, prompt injection, and the challenges of using AWS Lambda with chat GPT. Learn about the adoption of Bedrock in different industries and the benefits of response streams in both Lambda and Bedrock.

In this episode, I caught up with Ben Kehoe, who is an AWS Serverless Hero and one of the earliest adopters of serverless technologies.In a wide-ranging conversation, we discussed many topics around serverless and AI, including:The natural evolution of marketing terms and the need to focus on specific functional characteristics rather than defending the term. For example, including of arguing about what "serverless" means, we should instead talk about "pay-per-use".AWS should focus DX around the core service (e.g. CloudFormation) rather than trying to find client-side solutions by adding workarounds in SAM, CDK, etc.These client-side answers have a higher Total Cost of Ownership (TCO). Developers often don’t see the increased TCO they are taking on, but when things break, it’s a problem.Developers put too much emphasis on author time benefits and not enough on runtime and operational time costs. They should be more thoughtful about the operational time cost.The “infrastructure from code” movement is taking burdens off the developer but leaving them with the developer’s business, which is a bad thing.Developers often have a hard time separating delivering business value vs. coding.As an industry, a flawed narrative has emerged that developers are somehow special within an organisation and that it’s OK for them to ignore their responsibilities to security if there is friction in the process.Ai has the potential to impede human growth as the current AI systems are not designed to generate new ideas and challenge the status quo. “an AI generator that is trained on modernist art would never invent post-modernism”.Links from the episode:The meaning(lessness) of serverlessServerless is a state of mindThe serverless spectrumEp16 - Serverless at iRobot with Ben KehoeFor more stories about real-world use of serverless technologies, please follow me on Twitter as @theburningmonk and subscribe to this podcast.Want to step up your AWS game and learn how to build production-ready serverless applications? Check out my upcoming workshops and I will teach you everything I know.Opening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

In this episode, I spoke with Eduard Bargues, who is a Principal Engineer at Ohpen, a cloud-native open banking platform.We talked about Ohpen's migration from EC2 to a mix of Fargate and Lambda, and along the way, we touched on many topics:migration patternsthe "serverless-first" mindsethow to choose when to use Lambda vs Fargatewhy monolith functions are painful and should be avoidedrecommendations and pitfallswhy you should build a DX teamWe also talked about how being cloud-agnostic makes you use the cloud in a way that is inefficient and creates many layers of unnecessary abstraction layers in your architecture, which Eduard calls "legacy lock-in"!We talked about their event-driven architecture and how they are using an in-house Event Broker to add FIFO support (which EventBridge doesn't support) for some subscribers. It's similar to what Luc van Donkersgoed talked about in episode 68. We talked about their EventBridge topology and how they manage over 200 AWS accounts, and finally, what are the biggest shortcomings with EventBridge right now.Links from the episode:Job openings at OhpenEpisode 68 with PostNLEpisode 73 with NNLuc van Donkersgoed's Twitter and LinkedInThe decoupled invocation patternFor more stories about real-world use of serverless technologies, please follow me on Twitter as @theburningmonk and subscribe to this podcast.Want to step up your AWS game and learn how to build production-ready serverless applications? Check out my upcoming workshops and I will teach you everything I know.Opening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

In this episode, I spoke with Luca Mezzalira, who is the author of "Building Micro-Frontends" by O'Reilly, and a Principal Specialist Solutions Architect at AWS focusing on serverless.We talked about micro-frontends and dived deep into his experience with them at both DAZN and AWS. We got a high-level overview of micro-frontends: what it is, how they work, and who should adopt micro-frontends. And why it worked out so well at DAZN, which had the challenge of supporting more than a dozen of devices including set-top boxes with custom versions of HTML!We were also able to dive deep into some technical topics, such as how to manage the shared user state (e.g. credentials and session); how to build micro-frontends from a CI/CD pipeline POV; and how to maintain a consistent styling and layout across all your micro-frontend pages.Links from the episode:The frontend-discovery-service open-source project in AWS LabsLuca's Microfrontends bookFor more stories about real-world use of serverless technologies, please follow me on Twitter as @theburningmonk and subscribe to this podcast.Want to step up your AWS game and learn how to build production-ready serverless applications? Check out my upcoming workshops and I will teach you everything I know.Opening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0

In this episode, I spoke with Eric Peterson who is the founder and CTO of CloudZero, a leading vendor that helps AWS customers reign in their AWS spending by revealing hidden savings and preventing costly mistakes.We discussed a range of topics, including common mistakes AWS users make and how CloudZero once saved a customer potentially a million-dollar on CDN bills by alerting them of spending abnormalities early.We also talked about the trade-off between having cloud cost controls and the cost overhead they add. We talked about FinOps and how CloudZero can understand the unit cost of their own systems and how much each customer costs them because of the pay-per-use pricing model of serverless technologies.And we discussed the wider economic trends in 2023 and how that might impact companies going forward and how it might change their views on serverless and the cloud more broadly.Links from the episode:Erik's Twitter, Linkedin and personal blogBuilding a fully serverless insurance company with Joe EmisonFor more stories about real-world use of serverless technologies, please follow me on Twitter as @theburningmonk and subscribe to this podcast.Want to step up your AWS game and learn how to build production-ready serverless applications? Check out my upcoming workshops and I will teach you everything I know.Opening theme song:Cheery Monday by Kevin MacLeodLink: https://incompetech.filmmusic.io/song/3495-cheery-mondayLicense: http://creativecommons.org/licenses/by/4.0