Señors at Scale - Software Engineering & Tech Leadership

Dan Neciu
  • 0.0 (0)
  • Technology
  • Updated Weekly

A software engineering podcast for senior developers, staff engineers, and tech leads who build and scale systems in production. Hosted by Neciu Dan, Señors @ Scale features deep, technical conversations with engineering leaders from companies like Google, AWS, Microsoft, Cloudflare, Datadog, and Snyk. Every week, we sit down with Staff Engineers, Principal Engineers, and technical leaders to unpack the real challenges of frontend architecture, micro frontends, React and Vue at scale, design systems, security, reliability, and technical leadership. No fluff, no surface-level takes. Just hard-

  1. 6d ago

    Agentic Systems at Scale with Vilva Athiban (Omio) | MCP, Orchestration Agents, LLM Cost Optimization

    What does an agentic AI system look like when it has to work 100% of the time for millions of travelers, not just in a demo? Vilva Athiban is Lead AI Engineer at Omio, one of Europe's leading travel platforms. In June 2024, Omio picked him and one other engineer to start an AI team reporting directly to the CEO and CTO. He had no ML background. All he knew was how to call an LLM API. Since then he has built Omio's internal AI platform, the MCP integrations that connect agents to company systems, and Omio.ai, a consumer-facing agent product built entirely on MCP with no agent framework. In this episode Vilva walks through the full evolution of the architecture: LangChain, then LangGraph for control, then no framework at all. He explains the difference between supervisor and orchestration agents, why bigger models turned out cheaper than mini models at scale, how they cap tool calls and cache prompts to control cost, why guardrails ended up with the legal team, and why he thinks loop engineering will die in a few months. We also cover harnesses and context compaction, whether RAG is dead, how to pick models in 2026, fine-tuning, and what a JavaScript engineer should learn to move into AI. Key Topics: - From JavaScript engineer to Lead AI Engineer with no ML background - LangChain vs LangGraph, and why Omio's backend is now 100% MCP - Supervisor agents vs orchestration agents - Bigger models with medium thinking were cheaper than mini models - Capping tool calls, summarizing tool responses, prompt caching - Guardrails are subjective: the whale hunting debate - The agent is only as powerful as the tools you give it - What MCP is and why it replaced RAG for Omio - Harnesses, memory, and context compaction in Claude Code - Loop engineering: a sugar-coated while loop - How to choose models in 2026 (big, mini, nano) - Why fine-tuning is coming back in 2027 - Are developers cooked? The autocomplete that writes 15 files - How JavaScript engineers can become AI engineers - The token bubble, budgets, and the move to open source models CONNECT WITH VILVA 💼 LinkedIn: https://www.linkedin.com/in/vilvaathiban/ 🐦 X: https://x.com/vilvaathibanpb 🐙 GitHub: https://github.com/vilvaathibanpb 🌐 Website: https://www.vilvaathiban.com/ FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📧 Newsletter: https://neciudan.dev/subscribe 💼 Dan on LinkedIn: https://www.linkedin.com/in/neciudan 💼 Señors at Scale on LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Señors at Scale on Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan on Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES - Anthropic Learn: https://www.anthropic.com/learn - Model Context Protocol: https://modelcontextprotocol.io - Omio: https://www.omio.com #AIEngineering #MCP #AgenticAI #LLM #LangChain #LangGraph #SoftwareEngineering #AIAgents 💬 Have you hit the point where a bigger model was cheaper than a mini model in production? Tell me in the comments.

  2. Sep 7

    Decentralized Social Media at Scale with Dan Abramov (Vercel) | AT Protocol, Bluesky, Redux, React

    What if every app's data lived in one shared, open database that no company owns? Dan Abramov says that future already exists, and he spent two years building it. Season 3 of Señors at Scale opens with one of the most recognizable names in frontend. Dan Abramov is the co-creator of Redux, spent nearly a decade on the React core team at Meta, writes overreacted.io, and spent the last couple of years as an engineer at Bluesky working on the app and the AT Protocol. He is now at Vercel working on Next.js. In this episode Dan explains AT Protocol from first principles: why it is a specification and not a framework, how it splits hosting from apps, why your followers are not your data, and why he thinks Mastodon is a feudal kingdom passing messages on horses. He walks through migrating his own hosting to Eurosky, what it costs to run your own data server, and where a developer should start building. Then we get into Next.js, Turbopack vs Vite, the React Compiler, and why the creator of Redux has not needed a state manager in five years. Key Topics: - What AT Protocol is and why it is being brought into the IETF - The "everything folder": one repository per user, shared across all apps - AT Proto vs Mastodon and why message passing between instances does not scale - How a post from someone with 10 million followers actually flows through the network - Hosting, relays, PLC identity and the three things in your "passport" - How the Bluesky app itself is built (Go backend, React Native web frontend) - How companies make money when users own their data - Migrating your data hosting to Eurosky and the paper cuts along the way - did:plc vs did:web and what happens if you lose your domain - Why followers, likes and reposts are just links between JSON records - Running your own PDS on a VPS or Cloudflare Workers for close to free - Getting started: PDSLS, Tap, Constellation and building without a backend - Apps Dan wants to see on AT Proto and why cloning existing products is not enough - Vercel, Next.js and what Dan wants to fix - Turbopack vs Vite and the "no bundling in dev" bet - React Compiler as fine grained reactivity - Redux as a to-do list: where state should actually live - Book recommendation: Terence Tao's Analysis I (and its Lean companion) CONNECT WITH DAN 🦋 Bluesky: https://bsky.app/profile/danabra.mov 🐙 GitHub: https://github.com/gaearon 🌐 Blog: https://overreacted.io 🌐 Website: https://danabra.mov FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📧 Newsletter: https://neciudan.dev/subscribe 💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan 💼 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Show Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan's Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES - AT Protocol: https://atproto.com - "There are no instances in AT Proto" (Dan's blog post): https://overreacted.io - PDSLS repository explorer: https://pdsls.dev - Tangled (GitHub on AT Proto): https://tangled.org - Terence Tao, Analysis I Lean companion: https://github.com/teorth/analysis #ATProtocol #Bluesky #DanAbramov #React #Redux #NextJS #Decentralized #OpenWeb #WebDevelopment #SoftwareEngineering #SeñorsAtScale 💬 Would you move your data hosting off Bluesky's default servers? Or does owning your own JSON not matter to you? Let me know below.

  3. Aug 1

    Accessibility at Scale with Craig Abbott (TetraLogical) | WCAG, Screen Readers, Design Systems

    Your product passes every automated accessibility check. Does that mean anyone can actually use it? Craig Abbott is a Principal Accessibility Specialist at TetraLogical and the former Head of Accessibility at the UK's Department for Work and Pensions, one of the largest government departments in the country, where he built a dedicated accessibility practice from nothing and open sourced the DWP Accessibility Manual. He has over 15 years in user centred design, previously working as a designer and design manager, and has led accessibility work across the public sector and at Elastic. In this episode we get into what "sustainable accessibility" actually means, why culture is harder to change than capability, and how to stop accessibility from walking out the door when the one passionate person leaves. Craig breaks down his three C's framework (compliance, culture, capability), the screen reader testing stack he actually recommends, and what you can realistically automate in a CI pipeline with Playwright, Cucumber and Guidepup. We also get into the uncomfortable stuff. Automated tools only catch around 40% of issues. A page can pass WCAG with two pixel text and invisible links. And with 95.8% of the top million websites failing automated checks, the LLMs writing your front end code were trained on all of it. Craig is autistic and ADHD, and he closes on cognitive accessibility, the WCAG AAA gap, and why designing for neurodivergent users is still mostly ignored. Key Topics: - What sustainable accessibility means and why third party audits alone don't get you there - Building an accessibility practice from scratch inside a 100,000 person organisation - The three C's: compliance, culture and capability, and why culture is the hardest - Embedding accessibility so it survives when the champion leaves - JAWS vs NVDA vs VoiceOver vs TalkBack, and why the "worse" reader is the better test - Running screen readers in VMs and BrowserStack without expensive licences - Accessibility acceptance tests in CI with Playwright, Cucumber and Guidepup - Why compliant does not mean usable, with two pixel fonts and invisible links as proof - Shifting accessibility left into design and research, not throwing it at developers - Why you cannot detect screen reader users in your analytics - The three highest impact things: automated tools, keyboard only, structured content - Why CSS and design systems aren't opinionated enough - AI-generated front-end code and the 95.8% problem - Whether AI browsing replaces screen readers, and why LLM interfaces are inaccessible too - Designing for neurodivergence, WCAG AAA, and the COGA guidance CONNECT WITH CRAIG 💼 LinkedIn: https://www.linkedin.com/in/abbott567/ 🐦 X: https://x.com/abbott567 🦋 Bluesky: https://bsky.app/profile/craigabbott.bsky.social 🌐 Website: https://www.craigabbott.co.uk FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan 🏢 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Show Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan's Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES TetraLogical: https://tetralogical.com WebAIM Million report: https://webaim.org/projects/million/ W3C Cognitive Accessibility (COGA): https://www.w3.org/WAI/cognitive/ Guidepup: https://www.guidepup.dev/ axe-con: https://www.deque.com/axe-con/ #accessibility #a11y #webdevelopment #wcag #frontend #inclusivedesign #softwareengineering #neurodiversity #designsystems #senorsatscale 💬 Has your team ever shipped something fully compliant that users still couldn't use? Tell me about it in the comments.

  4. Jul 26

    Versatility at Scale with Carmen Huidobro (CTO, Incredible Bee) | 17 Years Freelancing, Automation, Ethical AI

    Is versatility the skill that actually survives the AI shift? In this episode of Señors at Scale, I sit down with Carmen Huidobro, CTO at Incredible Bee in Vienna. Carmen has spent 17 years in tech, almost all of it freelancing, working across Objective-C, Ruby on Rails, the web, mobile, hardware, and even ABAP in an SAP consultancy. Along the way, she spent five years in developer relations and developer education, and now leads engineering on long-lived macOS products like Renamer. Her argument is that the debate about generalists versus specialists misses the point. What matters is being an expert at adapting: recognizing patterns, carrying solutions across stacks, and dropping into unfamiliar environments without bringing chaos with you. Niches are not going away, but the durable skill is the one that moves between them. We also get into the part most teams get wrong: how to ship AI without betraying the people who use your software. Carmen walks through adding a local LLM to a twenty-year-old Mac app using Mistral and a llama.cpp runtime, why the model downloads on demand instead of shipping in the binary, and why it is opt-in rather than opt-out. Plus, the refugee hackathon app she built in 2016 that the City of Vienna still runs today, why she will not take money from junior developers for bootcamps, and what she calls "white guy-driven development". Key Topics: - Why versatility beats both generalism and specialization - Seventeen years of freelancing across stacks, platforms, and industries - SAP, ABAP, and why niche expertise never disappears - Building an ethical AI toolkit: Mistral, llama.cpp, and on-device inference - Adding an LLM to a legacy macOS app without breaking user trust - Opt in over opt out, and why you should never take features away - Deciding what should be automated and what should not - Pushing back when investors demand AI in the product - Where2Help: the 2016 refugee hackathon project in Vienna still runs - Why the junior developer market broke and what bootcamps can still fix - Stop telling women they have imposter syndrome - What AI will never teach you about solving problems CONNECT WITH CARMEN 🌐 Website: https://carmenh.dev 🐦 X: https://x.com/hola_soy_milk 🐙 GitHub: https://github.com/hola-soy-milk 🐘 Mastodon: https://hola-soy-milk.online/ FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 💼 My LinkedIn: https://www.linkedin.com/in/neciudan 🏢 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Show Instagram: https://www.instagram.com/senorsatscale/ 📸 My Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES Incredible Bee: https://incrediblebee.com Bad Website Club: https://badwebsite.club Distribute Aid: https://distributeaid.org #Versatility #Freelancing #Automation #EthicalAI #LocalLLM #Mistral #macOS #OpenSource #SoftwareEngineering #CTO #Podcast #SenorsAtScale 💬 Are you betting on going deep or going wide right now? Tell me in the comments.

  5. Jul 19

    CI/CD at Scale with Marko Gaćeša (Semaphore) | Semaphore, Product Engineering, Agent Native

    Is testing, not coding, the real bottleneck of the AI era? Marko Gaćeša is Head of Product at Semaphore, the agent native CI/CD platform built for coding agents: fast, affordable, enterprise-ready, and fully open source. A serial entrepreneur and technologist, Marko founded Dry Tools, served as Chief Product Officer at Alchemy Cloud, and has spent his career building high-quality software across domains such as medical equipment and industrial automation before entering the crowded CI/CD market. In this episode, Marko breaks down what it takes to build a product for developers: why flaky tests are the worst enemy of fast delivery, how SemAI uses skills and agents to automate everything from GitHub Actions migration to flaky test analysis, and why the future developer experience lives inside your coding agent, not in a browser tab. We also go deep on the product side: running experiments with feature flags, pricing a CI/CD platform when a "minute of CI" means something different everywhere, product-led growth versus B2B sales, and structuring lean product teams when your engineers are the domain experts. Key Topics: - What Semaphore does and why CI/CD matters - Testing is the biggest bottleneck in the age of coding agents - What "agent native" CI/CD really means - Flaky tests and how SemAI helps agents debug them - Migrating from GitHub Actions automatically - How Semaphore achieves faster and cheaper builds - Running product experiments on infrastructure tools - Pricing strategy in the CI/CD market - Should AI agents trigger deployments? - Product team structure in developer tools - Book recommendations for product-minded engineers CONNECT WITH MARKO & SEMAPHORE 💼 Marko on LinkedIn: https://www.linkedin.com/in/markogacesa/ 🌐 Semaphore: https://semaphore.io/ 🐦 Semaphore on X: https://x.com/semaphoreci 🐙 Semaphore on GitHub: https://github.com/semaphoreio 💼 Semaphore on LinkedIn: https://www.linkedin.com/company/semaphoreci FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 📸 Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan's Instagram: https://www.instagram.com/neciudev 💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan 💼 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ ADDITIONAL RESOURCES - Semaphore open source repo: https://github.com/semaphoreio/semaphore/ - Semaphore CI/CD benchmark: https://semaphore.io/semaphore-ci-cd-benchmark-performance-and-cost-analysis - Semaphore docs: https://docs.semaphore.io/ - Books mentioned: Hug Your Haters by Jay Baer, plus Inspired, Empowered, and Transformed by Marty Cagan and the Silicon Valley Product Group #CICD #Semaphore #ProductEngineering #AgentNative #DevTools #DeveloperExperience #AIAgents #SoftwareEngineering 💬 Would you let an AI agent trigger your production deployments? Tell us in the comments.

  6. Jul 12

    AI Harness at Scale with Maxim Salnikov (Microsoft) | GitHub Copilot, Agent Ops, Context Engineering

    Everyone can buy the same GitHub Copilot licenses. So what actually makes a company AI native? In this episode of Señors at Scale, Dan Neciu sits down with Maxim Salnikov, AI Dev Tools Solution Engineer at Microsoft, where he leads AI native development enablement for over 100 enterprise customers of Microsoft and GitHub in Norway. Maxim has been building for the web since the late 90s and now spends his days inside real enterprise dev teams, watching how AI adoption succeeds and fails across finance, energy, agriculture, and pure software companies. Maxim breaks down why AI adoption is a change management problem, not a technology problem, and why buying licenses is the easy part. We get into the anatomy of an AI harness, the internal layer that ships with tools like Copilot and Claude Code, and the external layer that successful companies build on top for their own codebases. He also makes the case for two new job titles emerging inside dev teams: context engineers and agent ops. Key Topics: - Why startups outpace enterprises in AI adoption, and how to close the gap - Handling AI skeptics: why one bad experience in 2023 shouldn't be the final verdict - Code as a secondary artifact: the harness is what you actually build - Internal vs external harness layers, and how the VS Code team builds its own tools - Context engineers and agent ops: the new roles appearing in team topology - The AI superheroes anti-pattern and why it doesn't scale - Managing agent skills like dependencies with APM (Agent Package Manager) - How to become an AI native engineer: orchestration over one-to-one chat CONNECT WITH MAXIM 💼 LinkedIn: https://www.linkedin.com/in/webmax/ 🐦 X: https://x.com/webmaxru 🐙 GitHub: https://github.com/webmaxru 🌐 Sessionize: https://sessionize.com/maxim-salnikov/ FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan 🏢 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan's Instagram: https://www.instagram.com/neciudev ADDITIONAL RESOURCES 📖 Agentic SDLC Handbook (free book mentioned in the episode) 📦 APM (Agent Package Manager) #AINativeDevelopment #GitHubCopilot #ContextEngineering #AIHarness #SoftwareEngineering #Microsoft #AgenticAI #DeveloperProductivity 💬 Is your team building its own harness on top of AI dev tools, or running them out of the box? Drop a comment 👇

  7. Jul 4

    URL State at Scale with François Best | nuqs, React, TypeScript

    What if your app's state didn't live in memory, but in the URL? François Best is a freelance web developer, open sourcer, and founder of 47ng, where he builds web applications with a privacy-first approach. He is the creator of nuqs, the type-safe search param state manager for React with over 10,000 GitHub stars, used by companies like Sentry, Supabase, Vercel, and Clerk. In this episode, François shares the full story of nuqs, from a woodworking calculator he built during the pandemic to a framework-agnostic library supporting Next.js, React Router, Remix, and TanStack. We dig into the internals: the async update queue, browser rate limits on the History API, throttling and debouncing, render count budgets, and the end-to-end test suite that keeps it all stable across frameworks. We also go beyond the library. François breaks down the recent npm supply chain attacks, explains how staged publishing and reproducible builds protect maintainers, shares how AI agents changed his workflow (including a Cypress to Playwright migration done by Claude Code on Christmas Day), and makes the case for running local AI models when your code can't leave the building. Key Topics: - What nuqs is and the origin story behind it - URL as the source of truth and the async scheduling engine - Browser rate limits, throttling, and debounce semantics - Framework adapters: Next.js, React Router, Remix, TanStack - End-to-end testing across frameworks and render count budgets - What state belongs in the URL (and what doesn't) - Hidden nuqs features: loaders, serializers, URL keys, standard schema - The npm supply chain attacks and how to defend as a maintainer - AI agents in open source: migrations, RFC prototyping, slop PRs - Local AI models for privacy-conscious companies CONNECT WITH FRANÇOIS 🐙 GitHub: https://github.com/franky47 🦋 Bluesky: https://bsky.app/profile/francoisbest.com 🌐 Website: https://francoisbest.com FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📬 Newsletter: https://neciudan.dev/subscribe 📸 Instagram: https://www.instagram.com/senorsatscale/ 📸 Dan's Instagram: https://www.instagram.com/neciudev 💼 Dan's LinkedIn: https://www.linkedin.com/in/neciudan 💼 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ ADDITIONAL RESOURCES - nuqs: https://nuqs.dev - nuqs on GitHub: https://github.com/47ng/nuqs - TkDodo's blog: https://tkdodo.eu/blog - Aurora Scharff's blog: https://aurorascharff.no #React #nuqs #TypeScript #URLState #NextJS #OpenSource #WebDevelopment #SoftwareEngineering #SenorsAtScale 💬 What state do you keep in the URL, and have you ever rolled your own search param sync? Tell us in the comments.

  8. Jun 28

    WebMCP at Scale with Susanna Wong (Typeform, Google Developer Expert) | Browser AI, Offline Models, Agents

    What happens to frontend engineering when software stops being deterministic? In this episode of Señors at Scale, Dan Neciu sits down with Susanna Wong, Staff Software Engineer at Typeform and Google Developer Expert in Web Technologies since 2020. Susanna's path is anything but typical: she studied engineering in the US, spent close to seven years as an architect and computational designer at firms like Arup, then pivoted back into software, building Kedro and Kedro-Viz at QuantumBlack before moving into AI features at V7 and now Typeform. Susanna breaks down the shift from training your own ML models to building on top of LLMs, what Web AI and offline browser models really offer, and the cross-browser hurdles still blocking local-first AI. She also takes us inside Typeform's multi-agent system, how the team thinks about evals for non-deterministic features, and why WebMCP could change how all of us build websites. Key Topics: - From architecture and computational design to frontend AI - Building Kedro and Kedro-Viz at QuantumBlack - Why training your own models gave way to the LLM baseline - Web AI, offline models, and on-device inference - The cross-browser problem with local-first AI - Typeform's multi-agent system: LangChain, A2A, MCP tools - Evals as unit tests for non-deterministic AI features - Offline vs online evals and tracking drift in production - WebMCP: bringing the web to agents 💼 LinkedIn: https://www.linkedin.com/in/susanna-wong-london/ 🐦 X: https://x.com/studioswong 🌐 Talks: https://www.youtube.com/watch?v=JcLZGfXKk8c FOLLOW & SUBSCRIBE 🎙️ Podcast: https://neciudan.dev/senors-at-scale 📩 Newsletter: https://neciudan.dev/subscribe 💼 Show LinkedIn: https://www.linkedin.com/company/senors-scale/ 📸 Show 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 - Kedro: https://kedro.org - WebMCP / React Paris talk on the agent web - Typeform AI: https://www.typeform.com #WebAI #WebMCP #SoftwareEngineering #Frontend #AIAgents #Evals #StaffEngineer #Typeform #MachineLearning #Podcast #SenorsAtScale 💬 Are you building with offline browser models yet, or still calling the API for everything? Let me know below.

See All (50)

About

A software engineering podcast for senior developers, staff engineers, and tech leads who build and scale systems in production. Hosted by Neciu Dan, Señors @ Scale features deep, technical conversations with engineering leaders from companies like Google, AWS, Microsoft, Cloudflare, Datadog, and Snyk. Every week, we sit down with Staff Engineers, Principal Engineers, and technical leaders to unpack the real challenges of frontend architecture, micro frontends, React and Vue at scale, design systems, security, reliability, and technical leadership. No fluff, no surface-level takes. Just hard-

Information

You Might Also Like