Señors at Scale - Software Engineering & Tech Leadership

TanStack Query at Scale with Dominik Dorfmeister (TkDodo) | Open Source, Knip, Sentry Design System

What does it actually feel like to maintain a library used by millions of developers every day? In this episode of Seniors at Scale, Dan sits down with Dominik Dorfmeister, better known as TkDodo, the creator and maintainer of TanStack Query and a software engineer at Sentry. Dominik has spent over a decade building frontend tooling and has become one of the most trusted voices in the React and TypeScript ecosystem. Dominik shares how he got into open source during the pandemic lockdowns, simply by answering questions in Discord, and how that habit grew into maintaining one of React's most widely adopted libraries. He talks candidly about the breaking change that went wrong, why major versions are "the pain of his existence," and what he has learned about shipping changes to a community that only shows up with feedback after release. The conversation also digs into his work at Sentry, where he used Knip to remove 28,000 lines of dead code, and where his team is building a new design system inside a 10 year old, million-line codebase. Key Topics: - Getting into open source by answering community questions - Becoming the maintainer of TanStack Query - Tracked queries and the first big performance feature - Why major version releases are so painful - The version 4 to 5 breaking change that went wrong - Epoch versioning as an alternative to semver - Using Knip to find and remove dead code - Building a design system in a large, legacy codebase - What is planned for TanStack Query version 6 Guest: Dominik Dorfmeister (TkDodo) 🌐 Blog: https://tkdodo.eu/blog 🐦 X: https://x.com/TkDodo 🐙 GitHub: https://github.com/TkDodo FOLLOW & SUBSCRIBE 💼 LinkedIn: https://www.linkedin.com/company/senors-scale/ 🌐 Podcast: https://neciudan.dev/senors-at-scale 📨 Newsletter: https://neciudan.dev/subscribe 📸 Instagram: https://www.instagram.com/senorsatscale/ Connect with Dan: 💼 LinkedIn: https://www.linkedin.com/in/neciudan 📸 Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES TanStack Query: https://tanstack.com/query Knip: https://knip.dev Sentry: https://sentry.io TkDodo's blog: https://tkdodo.eu/blog #ReactJS #TanStackQuery #OpenSource #FrontendDevelopment #SoftwareEngineering #WebDevelopment #TypeScript #Sentry 💬 What is your take: should libraries do fewer, bigger major versions, or more frequent, smaller ones? Let us know in the comments.

Performance Engineering at Canva with Den Odell (Staff Engineer & Manning Author)

What happens to "edge cases" when your product serves 250 million people every month? In this episode of Señors at Scale, I'm joined by Den Odell, Staff Software Engineer at Canva and author of "Performance Engineering in Practice" (Manning, 2026). Den works on Canva's Pro Design group, building inside one of the largest React/TypeScript codebases on the planet, serving 250M+ monthly users across 190+ countries. Before Canva, Den spent 9 years at Volvo Cars architecting their Offer Selector and Car Configurator (powering vehicle purchases across 70+ markets), and 13 years at AKQA leading global frontend engineering for Nike, MINI, and Nokia. He's authored three books, the latest of which introduces the Fast by Default framework, a methodology for treating performance as a design decision from day one rather than a panic fix at the end. We get into how Canva ships safely at a planetary scale (feature flags, dogfooding, geofenced rollouts, test parties), the protobuf-based RPC layer powering their frontend/backend communication, async-first culture across global timezones, and why most teams are stuck in what Den calls the Performance Decay Cycle. Key Topics: * Why bugs scale with your user base (and what to do about it) * Canva's release pipeline: staff → beta → geofenced regions → world * Test parties, dogfooding, and catching weirdness before users do * Protogen, CDF, and how Canva moves data between frontend and backend * Operational transforms for real-time collaboration * The Performance Decay Cycle and why "performance sprints" are broken * Fast by Default: making speed everyone's responsibility, not just engineering * Perceived performance, AI loading states, and the Pac-Man tape loader lesson * Async work culture when "the sun never sets on the Canva Empire" * Building for crazy big goals (what does Canva look like at 1 billion users?) GUEST SOCIALS💼 LinkedIn: https://www.linkedin.com/in/denodell🐦 X/Twitter: https://x.com/denodell🐙 GitHub: https://github.com/denodell🌐 Website: https://denodell.com FOLLOW & SUBSCRIBE🌐 Podcast: https://neciudan.dev/senors-at-scale📧 Newsletter: https://neciudan.dev/subscribe💼 Show LinkedIn: https://www.linkedin.com/company/senors-scale/📸 Instagram: https://www.instagram.com/senorsatscale/💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan📸 Dan's Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES📘 Performance Engineering in Practice (Manning, 2026): https://www.manning.com/books/performance-engineering-in-practice🔧 Code listings on GitHub: https://github.com/denodell/performance-engineering-in-practice✍️ Den's blog: https://denodell.com/blog📕 The Product-Minded Engineer (Drew Hoskins, O'Reilly): mentioned in episode #Frontend #Canva #FeatureFlags #Dogfooding #PerformanceEngineering #SoftwareEngineering #WebDevelopment #StaffEngineer #ReactJS #TypeScript #Podcast #SenorsAtScale 💬 What's the worst "edge case turned major incident" you've shipped to production? Drop it in the comments.

Frontend at Meta with Evyatar Alush | Hack, Flow, Sapling, Open Source at Scale

What does engineering at Meta actually look like from the inside? Spoiler: almost nothing you know from outside applies. In this episode, Dan sits down with Evyatar Alush, Software Engineer at Meta in Tel Aviv and the creator of EmojiPicker React (600K+ weekly downloads) and Vest. Evyatar's journey is one of the most unusual on the show: no degree, no high school diploma, learned JavaScript on Code Academy during military night shifts in a server room, then talked his way into Fiverr, scaled to Front End Platform Lead, and got recruited into Facebook in 2019. We get into what it's actually like inside Meta's frontend infrastructure: Hack instead of PHP, Flow instead of TypeScript, Relay instead of Apollo, Sapling instead of Git, stacked diffs instead of pull requests, and a custom everything (testing frameworks, ORMs, dev servers, data centers). We also cover his open source philosophy, why he builds his own libraries instead of pulling dependencies, the supply chain risks of modern npm, and how AI-assisted code is reshaping open source maintainer work. Key Topics: - Learning to code on military night shifts with zero CS background - Joining Fiverr with one year of experience and bluffing through the interview - Building Fiverr's notification system, in-app inbox, and toast library - Creating EmojiPicker React from a Fiverr internal tool - The "Unmask" manifesto and starting Fiverr's frontend infrastructure team - Designing the Front Ants team by faking the trappings of a real team - Building micro-frontends that bridge a Ruby on Rails monolith and React - Saying no to Facebook on the first email - Interviewing at Meta in London (and the Dan Abramov interview) - The Calibra/Diem crypto wallet team during COVID - Hack vs PHP, Flow vs TypeScript, Relay vs Apollo, Sapling vs Git - Stacked diffs and why ex-Meta engineers miss them - Why "move fast and break things" is dead at Meta - Code review, dev mod servers, and end-to-end testing at scale - Open source maintenance in the AI era and Cursor-generated PRs - Why he owns the "context" package on npm GUEST: Evyatar Alush 💼 LinkedIn: https://www.linkedin.com/in/evyataralush-5b760866 🐙 GitHub: https://github.com/ealush 🌐 EmojiPicker React: https://github.com/ealush/emoji-picker-react 🌐 Vest: https://vestjs.dev FOLLOW & SUBSCRIBE 📸 Instagram: https://www.instagram.com/senorsatscale/ 📸 Instagram: https://www.instagram.com/neciudev 🎙 Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 💼 LinkedIn: https://www.linkedin.com/in/neciudan 💼 LinkedIn: https://www.linkedin.com/company/senors-scale/ ADDITIONAL RESOURCES - EmojiPicker React: https://www.npmjs.com/package/emoji-picker-react - Vest (form validation): https://vestjs.dev - Sapling (Meta's source control): https://sapling-scm.com - The Hack language: https://hacklang.org - Flow: https://flow.org - Relay: https://relay.dev - The Design of Everyday Things by Don Norman - Never Split the Difference by Chris Voss #Meta #Facebook #Frontend #ReactJS #HackLang #Flow #Relay #Sapling #StackedDiffs #OpenSource #EmojiPickerReact #Vest #SoftwareEngineering #SenorsAtScale 💬 What's your take on Meta's "everything in-house" engineering culture? Would you rather work with familiar tools or relearn engineering from scratch for better internal infrastructure?

React Native at Scale with Kadi Kraman, Software Developer at Expo | Mobile Development, EAS, OTA Updates

What does it actually take to build production React Native apps in 2026, and where does Expo fit in? In this episode of Señors at Scale, Dan sits down with Kadi Kraman, software developer at Expo, who has spent over six years in the React Native ecosystem, wearing every hat from IC to director. Kadi shares the story of how she went from writing C++ in a maths degree to becoming one of the early React Native engineers at Formidable, and eventually joining Expo to work on the platform itself. We dig into what makes React Native genuinely competitive with native iOS and Android development today, why Expo Go is now just for prototyping, how EAS workflows and fingerprint-based repacks dramatically speed up CI, the real story on OTA updates (and where the legal gray area sits), and what's still missing from the ecosystem. Kadi also gives a rare look at the new Expo agent for vibe-coding mobile apps, the case for React Native brownfield, and her honest take on Lynx as competition. Key Topics: - Why React Native + Expo is faster than native Xcode/Android Studio workflows - The mental shift from web to native (display points, gestures, pixel density) - Expo Go vs development builds, and why the recommendation has changed - EAS workflows, repack jobs, and project fingerprints - React Native performance, list rendering, and the React Compiler - OTA updates: when to use them, when not to, and what the stores actually allow - Debugging strategies (expo-doctor, native logs, AI-assisted log analysis) - Brownfield React Native and embedding RN screens into existing native apps - Lynx, competition, and the future of cross-platform mobile - Career advice on imposter syndrome, applying anyway, and finding talk topics GUEST: Kadi Kraman, Software Developer at Expo 💼 LinkedIn: https://www.linkedin.com/in/kadikraman/ 🐦 Twitter/X: https://x.com/kadikraman 🐙 GitHub: https://github.com/kadikraman 🌐 Website: https://kadikraman.com/ FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📩 Newsletter: https://neciudan.dev/subscribe 💼 LinkedIn (Show): https://www.linkedin.com/company/senors-scale/ 💼 LinkedIn (Dan): https://www.linkedin.com/in/neciudan 📸 Instagram (Show): https://www.instagram.com/senorsatscale/ 📸 Instagram (Dan): https://www.instagram.com/neciudev ADDITIONAL RESOURCES - Expo: https://expo.dev/ - Kadi's "From Web to Native with React" blog post: https://expo.dev/blog - EAS Workflows: https://docs.expo.dev/eas-workflows/get-started/ - Expo Doctor: https://docs.expo.dev/more/expo-cli/#expo-doctor - Expo Fetch (streaming support): https://docs.expo.dev/versions/latest/sdk/expo/ #ReactNative #Expo #MobileDevelopment #JavaScript #iOS #Android #EAS #ExpoRouter #SoftwareEngineering #SeñorsAtScale 💬 What's your biggest pain point building React Native apps today, and have EAS workflows changed your CI setup?

AI at Scale with Nico Martin from Hugging Face | Transformers.js, Tokenizers, On-Device Inference

Can you really run state-of-the-art machine learning models directly in the browser, with no server, no API calls, and full privacy by default? In this episode, Nico Martin, Open Source Machine Learning Engineer at Hugging Face and Google Developer Expert in AI and Web Technologies, walks through how Transformers.js makes on-device AI a reality. Nico's journey is anything but conventional. He started as a ski and windsurf instructor, taught himself web development on the side, spent years as a freelancer (including five at a bank building e-banking front ends), and recently landed what he calls his dream job at Hugging Face. We unpack what Hugging Face actually is (the GitHub for machine learning), how Transformers.js brings the Python Transformers API to the browser, and the real engineering challenges of running models on whatever hardware your users happen to have. Nico explains quantization, ONNX as the standard for portable model architectures, the role of tokenizers, how text becomes tensors, and why WebGPU matters for running larger models client-side. We also dig into the bigger picture: privacy-preserving AI, the difference between open weights and truly open source models, agents and MCP, and what front-end developers should actually learn to stay relevant in an AI-first world. Key Topics: - What Hugging Face is and the role of the Hub, Transformers, and Diffusers - Transformers.js: bringing Python Transformers API to JavaScript and the browser - The biggest challenge of browser ML: running on unknown client hardware - Quantization explained (Q4, 4-bit vs 16/32-bit) and how it compresses models - ONNX and ONNX Runtime Web: the standard for portable model architectures - Open weights vs open source models and why the distinction matters - Tokenizers, token IDs, and why each model needs its own tokenizer - From text to tensors: pre-processing, inference, and post-processing - Text embeddings explained through a simple animal feature analogy - WebGPU and what it unlocks for in-browser inference - Agents, tool calling, MCP, and how context windows get consumed - Advice for developers who want to break into AI and ML engineering 🔗 FOLLOW NICO 💼 LinkedIn: https://www.linkedin.com/in/nicodotdev/ 🐦 X/Twitter: https://twitter.com/nic_o_martin 🦋 Bluesky: https://bsky.app/profile/nico.dev 🐙 GitHub: https://github.com/nico-martin 🌐 Website: https://nico.dev 🎙️ FOLLOW & SUBSCRIBE 📸 Instagram: https://www.instagram.com/senorsatscale/ 📸 Instagram: https://www.instagram.com/neciudev 🎙 Podcast URL: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 💼 LinkedIn: https://www.linkedin.com/in/neciudan 💼 LinkedIn: https://www.linkedin.com/company/senors-scale/ 📚 ADDITIONAL RESOURCES - Transformers.js: https://huggingface.co/docs/transformers.js - Hugging Face: https://huggingface.co - ONNX: https://onnx.ai - ONNX Runtime: https://onnxruntime.ai - WebGPU: https://www.w3.org/TR/webgpu/ - Utopia for Realists by Rutger Bregman #MachineLearning #AI #HuggingFace #TransformersJS #WebML #OnDeviceAI #WebGPU #ONNX #JavaScript #Frontend #WebDev #SenorsAtScale #OpenSource 💬 Would you trust on-device AI over cloud-based models for sensitive data? Share your thoughts in the comments!