why bit summit?

breakfast and coffee are available to get you ready for a day of valuable tech content.

each track (backend, frontend, data science) will open with a panel talk featuring all the speakers for that track. get to know these speakers, their challenges, and the valuable lessons you can expect to discover throughout the day.

descripton:

At MoneyBank Inc., we’d just made the leap to Kubernetes and GitOps—ready to trade our feature frenzy for stability and scale. CI/CD was fully automated, microservices gleamed, but somehow… features still crawled out the door. Weeks for a release. Teams stuck waiting on each other. Frustration brewing. Trust eroding.

Classic enterprise-meets-startup clash: we raced to automate every problem, but culture and compliance didn’t get the memo. Automation hit its limit. In this talk, I’ll dive into what comes next—ReleaseOps: GitOps for the people. Because some things can’t be automated… until they are.

about the speaker:

I always wanted to save the world. After leaving law school, I decided to work with computers instead. While in Web Dev, I started attending tech events, and soon fell in love with the community. In my roles as Consultant and DevRel, I combined technical knowledge with a focus on the human side.

Currently, I work as a Platform Engineer in Amsterdam and serve as Co-Chair for the CNCF TAG App Delivery. I also enjoy performing in local musical theater and standup comedy shows.

description:

Design systems have evolved from simple component libraries to sophisticated ecosystems that govern how products are built at scale. But as AI fundamentally reshapes the software development industry, design systems face their biggest transformation yet—are we building design systems for humans or feeding structured data to increasingly hungry LLMs?

In this case study, I’ll take you behind the scenes of my journey at Adobe, where I used Cursor AI to develop design system components for Adobe Express. From the initial excitement of 10x faster prototyping to the sobering reality that AI-generated components often fail basic accessibility standards, this talk exposes the messy truth about scaling design systems in an AI-first world.
Here’s what I learned: The promise was seductive: import Figma plugin code, let Cursor write the React components, ship faster than ever. The reality? Building inclusive, accessible, and truly reusable components requires human judgment that no AI can replicate—yet. When Cursor suggested a beautiful dropdown that was completely unusable for screen readers, I learned that speed without intentionality is just technical debt in disguise.

By the end of this session, you’ll understand how to harness AI’s genuine strengths while protecting the human-centered principles that make design systems sustainable. Because the question isn’t whether AI will transform design systems—it’s whether we’ll intentionally shape that transformation or let it happen to us.

The critical question we’ll tackle: If components are the structural foundation of design systems, are design systems becoming just sophisticated kibble for LLMs—and what does that mean for the humans who actually use our products?

about the speaker:

I am a senior engineer at Adobe and a global voice for digital equality, to learn how to ensure your AI solutions truly serve everyone. I shape global accessibility standards through influential roles at the W3C and the International Color Consortium (ICC), driving forward inclusive innovation worldwide.Oh, and I am definitely a coffee enthusiast!

description:

At Flo Health, we faced a classic scale problem: managing a vast ecosystem of content—stories, articles, widgets—localized into over 20 languages, all while navigating complex medical and legal review processes. Our biggest bottleneck? Medical reviews by qualified professionals who were both expensive and scarce. What used to take weeks of back-and-forth between medical experts, legal teams, designers, and engineers was killing our ability to serve users with timely, personalized content.

In this talk, I’ll share how we transformed our content operations using GenAI and a “factual judge” framework, turning medical experts from manual reviewers into meta-experts who create evaluation datasets and quality rules. The result? An agent-based system where content creators submit simple prompts and receive fully generated, medically-evaluable content—cutting our time-to-market from weeks to hours while maintaining the rigorous standards our regulated environment demands. I’ll show you the architecture, the challenges we faced, and how you can apply similar approaches in your own regulated environment.

about the speaker:

I lead teams serving millions of women around the world. I’m passionate about building high-performing organizations that actually ship meaningful products. Over the years, I’ve gone from hands-on engineering to founding a startup to leading globally distributed teams across continents. I specialize in scaling platform teams and developing leaders — but what drives me most is creating environments where people can do their best work while tackling complex technical challenges.

descripton:

Dropped into a FinTech venture inside one of Britain’s oldest banks, I encountered a complex challenge: fragmented systems, unclear ownership, and one simple question from leadership that no one could answer. That moment exposed critical operational gaps—and set off a full-scale transformation.

In this talk, I show how GitOps isn’t just buzzword, it’s a mindset shift. He’ll walk through how they restructured their Git repos, rebuilt trust in delivery, and embedded GitOps across tech and culture alike. From “The Awakening” to “The New Reality,” you’ll get real-world lessons, practical patterns, and a blueprint for leading change through GitOps. Whether you’re scaling systems or reshaping teams, this is how to turn chaos into control with confidence.

about the speaker:

I fix failed Kubernetes migrations. With over 50 successful CNCF transformations under my belt, I help enterprises go from “drowning in YAML” to “shipping daily” — in just 30 days. Known for my “learn by breaking” approach and pragmatic view of the CNCF landscape, I speak the uncomfortable truths about cloud-native adoption that vendors won’t tell you. Former Flux maintainer, current migration firefighter.

description:

Scaling from 100 to over 50,000 RPS in 7 years at adjoe taught us one critical lesson: with great scale comes great cloud bills. In a distributed system that’s always in flux, understanding system behavior isn’t just important, it’s essential for survival.

In this talk, I’ll walk you through our observability evolution. You’ll learn why our Elasticsearch bill forced us into a hybrid logging strategy, the pros and cons of our leap from CloudWatch to Prometheus, and the universal challenges you’ll face when scaling.

We’ll explore the three pillars of observability (and more). Find out hard-won lessons that can save you from your own scaling headaches.

about the speaker:

I’m a Principal Backend Developer at adjoe, bringing over 10 years of experience in building and scaling reliable systems. I work with a wide range of technologies to drive improvements across our product — from optimizing performance and speed to reducing infrastructure costs and ensuring long-term scalability.

What I enjoy most is diving deep into complex challenges, whether it’s fine-tuning system architecture, streamlining backend workflows, or collaborating across teams to deliver impactful features. For me, it’s all about creating efficient, maintainable solutions that move the product and the business forward.

description:

Learn how to perform real-time image processing directly in the browser using TensorFlow.js, Google’s powerful JavaScript library for machine learning. This talk introduces client-side AI techniques that eliminate server round-trips, enabling fast, private, and interactive image analysis using Web APIs and pre-trained or custom models.
Designed for developers with basic JavaScript, HTML, and CSS skills, this session offers practical insights into integrating machine learning into web apps. Whether you’re building tools for clients or experimenting creatively, you’ll gain hands-on experience and a solid foundation to start using Web ML in production.

about the speaker:

I’m a passionate developer focused on engineering technologies aligned with the latest trends. My journey in web development has been driven by my experiences over the years, and I love sharing my excitement with others. Currently, I’m thrilled to be working at ING. Outside of my professional life, I’m dedicated to my family, ensuring their happiness is a top priority. I also find joy in speaking at conferences, contributing to open-source projects, helping people in communities, recording video podcasts, and playing video games.

description: 

Every time you open an app or website, an auction takes place to decide which ad you see. Behind the scenes, algorithms compete at a massive scale to determine the bid price. In this game, the goal isn’t just to win; it’s to win efficiently. Why bid a dollar for an impression when you could secure it for 80 cents? That 20-cent difference, when multiplied across millions of bids, is the critical factor separating profit from loss.

In this talk, I’ll show how we frame this as a surplus maximization problem, using probabilistic modeling, statistical learning, and large-scale optimization. We’ll unpack the math that drives our models, how we turn raw data into real-time bid predictions, and the engineering challenges of running AI-driven decision systems in millisecond timeframes. Even if you’ve never worked in adtech, you’ll see how these ideas connect to broader machine learning and optimization problems you might face in finance, logistics, or other high-frequency domains.

about the speaker:

I’m a Senior Data Scientist at adjoe, where I work at the intersection of data, scale, and real-time decision-making in programmatic advertising. I build and deploy models for user value prediction, conversion forecasting, and large-scale real-time bidding. I’m passionate about turning complex data into actionable insights — and exploring how emerging technologies can push the boundaries of what’s possible in big data.

description: 
In today’s fast-evolving tech world, serverless architecture transforms how we build and manage scalable applications by moving the responsibility of creating, configuring, and managing resources away from the engineers to the cloud provider. In this session, we will discuss the fundamentals of serverless computing and explore how Terraform can be a powerful ally in managing serverless infrastructure. Whether you’re a developer, DevOps engineer, or IT leader of any seniority, this session will provide valuable insights into leveraging serverless computing and Terraform to build efficient, scalable, and manageable cloud solutions.

about the speaker:
I am a Brazilian senior software engineer with eight years of experience, currently residing in Berlin. With extensive expertise in web development, I hold an AWS Cloud Architect Associate certification and am deeply passionate about cloud architecture, full-stack development, and infrastructure as code. My specialties include designing and implementing microservices and serverless solutions, ranging from straightforward API integrations to intricate event-driven asynchronous architectures. Beyond my technical skills, I am committed to sharing knowledge through my tech articles and have a constant thirst for learning new technologies to keep my content fresh and informative.

description:

New or experienced gopher, you can expect to emerge with a strong understanding of the Go type system.

If you are a Gopher, the mention of the term Object-Oriented Programming might be triggering for you because Object Oriented represents everything that Gophers love to hate, from over-complicated design to unpredictable execution. But could Go be in fact OO? Experienced or new to Go, this talk is aimed to help you understand where Go fits within the realm of programming languages. Ronna will take you on a journey through the history of OO and the choices that were made when designing Go (including some lesser understood features), so that you can decide for yourself if Go is (or isn’) OO.”

about the speaker:

I’m a technologist, a Google Developer Expert for Go, a Women Who Go organizer, and a GoTime “unpopular opinion” hall of famer. After over 20 years in software development, I know I’m the sum of the opportunities I’ve been given—which is why I dedicate time to helping others. I’ve been crafting hands-on workshops in Go since 2017.

description:

The most talked-about news in the React community was probably the React Compiler. Everyone is looking forward to being saved from the re-render plague and never having to write useCallback/useMemo again. But are we truly there yet? Can the React Compiler actually achieve those things? Do we need to do anything other than turning it on and enjoying a re-render-free life?

I ran it on a few real-world apps to find the answer. Want to see the results? In this talk, we’ll take a look at what the React Compiler is, what problems it solves, and how it performs on simple code examples and real-world applications.

about the speaker:

I’m an experienced Frontend Architect and UI/UX Engineer with a strong focus on React, web performance, and scalable frontend architecture. I specialize in bridging the gap between design and engineering, creating high-quality user experiences that are both visually polished and technically sound.

As an educator, author, and speaker, I’m passionate about developer tools, intuitive UI/UX, and making complex concepts easy to understand through clear explanations and visual storytelling.

description:

At DLR Hamburg’s Institute of System Architectures in Aeronautics, we’re making that vision a reality. By connecting disciplines, partners, and tools through a unified digital thread, we’re pioneering data-driven, climate-neutral, and cost-efficient aircraft design.

At the core is CPACS, an open-source XML standard that captures the full aircraft definition—from geometry to systems—and serves as a single source of truth. Supporting tools like TiXI (for efficient data access), TiGL (for generating precise 3D models), and CPACS Creator (a no-code GUI for configuration) make it easy to build and iterate on designs.

All of this runs through RCE, our powerful automation platform that manages end-to-end engineering workflows—simulation, cost estimation, performance analysis, and more. The results are showcased in the Digital Hangar, an interactive portal where users can explore, analyze, and download parametric aircraft models.

about the speaker:

I am a scientist at the German Aerospace Center (DLR), where I work on data-driven engineering methods for next-generation aircraft systems. My research focuses on machine learning and uncertainty quantification in digital aerospace design, and I contributes to the development of open tools that help engineers make better decisions through simulation and automation.

description:

What does it take to prepare your backend for a traffic spike of 100,000 requests per second? In this session, we’ll dive into the real-world challenges and solutions for scaling AWS infrastructure to handle massive, unpredictable loads—like those during a Super Bowl ad. Learn how adjoe optimized DynamoDB, leveraged Auto Scaling, pre-warmed resources, and loadbalancer sharding to ensure stability under extremely heavy loads in short duration. We’ll share key lessons on observability, load testing, and architectural tweaks that made the difference between success and failure. Whether you’re preparing for a major event or just want to build a more resilient system, this talk will equip you with actionable insights to scale confidently.

about the speaker:

I enjoy improving existing cloud infrastructure and building new systems when needed. Since starting my journey in DevOps over four years ago, I’ve loved exploring new tools and technologies. And last but not least, automation is my passion.

lunch is on us. enjoy!

description:

In under six months, my team faced two major outages from the same root cause: a total power loss in our primary data center. The first caused widespread disruption. The second? Barely a blip.

This talk shares how we transformed our systems for resilience—covering real-world disaster recovery, distributed patterns in Go, and techniques for reliable, high-availability architecture. Learn how to build systems that stay up when everything else goes down.

about the speaker:

Senior SRE at Google, building and maintaining large-scale, highly available, and robust distributed systems. I’ve worked in software engineering for over nine years, with the majority of that time focused on writing Go. My experience spans both big-name companies like Google and Cloudflare, as well as fast-paced, hyper-growth startups across various industries—including finance, cybersecurity, and the public sector.

description:

JavaScript performance isn’t just about bundle size. In this session, we’ll break down the real cost of JS—from network transfer to execution and rendering—and how each step affects your users’ experience.

With DevTools demos, insights into how V8 really runs your code (think hidden classes and inline caches), and hands-on fixes like code-splitting and FastDOM, you’ll leave with practical strategies to make your React + TypeScript apps run faster—on any device.

Whether you’re debugging layout thrash or setting performance budgets, this talk will help you ship faster, smoother experiences your users will feel.

about the speaker:

I’m a front-end developer passionate about building smooth, user-friendly interfaces. I believe great UX starts with empathy—and maybe a few too many sticky notes. I thrive in the space between chaos and clean design, always chasing the simplest solution that actually works.

description:

Let’s dive into the innovative world of Help Experience — Vinted’s community support engineering domain. Discover how Vinted leverages a strategic backend system design paired with modern technology and Artificial Intelligence to handle thousands of member requests daily. We’ll explore the transformation from a “big ball of mud” to a well-architected and scalable backend solution. We’ll touch on Vespa vector search to streamline information retrieval and learn about our Temporal setup for Virtual Assistant, utilizing large language models (LLMs) to enhance customer interactions.

Finally, we’ll touch on how to empower thousands of  Support Agents to deliver exceptional help. Join us to uncover the secrets behind Vinted’s effective community support solutions.

about the speaker:

As the Vinted Second Hand Marketplace grows, I drive scaling Customer Support technology while keeping it simple and preserving a joyful developer experience. Here, I leverage my passion for Clean Code,
robust Software Architecture, and effective API Design. I enjoy sharing lessons learned from real-world engineering challenges and creating technology that turns user frustrations into positive experiences.

description:

As a tracking platform we’re getting a lot of events from various internal and external sources. All this information has not only to be processed but also stored for multiple years for analysis and reprocessing.

In this talk I want to share you the different detours we took in our scaling path and how we’ve implemented the current state of art solution for data at rest: the lakehouse. You will get detailed information about our design choices and how we’re utilizing the different data engines like Trino, Spark and Flink with Iceberg. Follow me into the rabbit hole of git-inspired data version control.

about the speaker:

I’m the director of technology at justtrack, a mobile app and game marketing analytics platform. For over 10 years I’m building and scaling our platform from handling a handful to billions of events per day.

With deep expertise in backend engineering and cloud-native infrastructure, I love utilizing Golang and open source technologies to shape high performance and cost efficient solutions. As the creator of the gosoline application framework, I’m trying to give something back to the community.

description:

Reddit’s feed was originally built to serve only posts. As product requirements grew, so did the need to support new types of content — like community recommendations, announcements, and interactive updates — directly within the feed.

In this talk, I’ll share how we extended the feed system to support these experiences in a controlled, scalable way. I’ll talk about the engineering challenges behind introducing new content types into a mature, high-traffic system — from ensuring reliability and performance to enabling targeted delivery and preserving the core user experience.

about the speaker:

I’m a software engineer with almost 20 years in tech, passionate about building systems that scale — and matter. I’ve architected platforms used by millions, led backend solutions for banks and global companies, and validated systems under serious load. But my real drive is connecting solid engineering with product vision. I mentor women in tech, write about architecture that works in practice, and believe engineers shouldn’t just ship features — we should shape what gets built.

description:

Coding AI agents often produce non‑deterministic outputs, which is a challenge for predictability (and quality). In this talk, I’ll go way beyond function-calling schemas and runtime validations to leveraging Typescript for more deterministic and reproducible results.

about the speaker:

I’m a frontend engineer, a technical author with 5+ books published and an avid reader. Since 2017, I’ve blogged extensively, with my blogs garnering well over 4 million views. I don’t promise to know everything, and love fiddling with new technologies.

description:

Recommender systems are a core part of many modern products — but building one that works well and respects the user in the context of notifications is a uniquely challenging problem. In this talk, I’ll walk through Reddit’s push notification recommendation system: how it works, what we’ve learned, and why the real-world tradeoffs often differ from textbook solutions.

From cold starts to user fatigue, content diversity to delivery timing — I’ll share the challenges that make push recommendations complex, and how we’re addressing them in practice.

about the speaker:

Senior Machine Learning Engineer at Reddit. I love building great products and user experiences with ML, and I’m passionate about data, experimentation, and all things tech.

description:

I am Lead Architect in Microsoft’s Azure Messaging team that builds and operates a fleet of hyper-scale messaging services, including Event Grid, Service Bus, Event Hubs, Stream Analytics and Microsoft Fabric Eventstreams. I represent Microsoft in messaging standardisation in OASIS (AMQP, MQTT) and CNCF (CloudEvents, xRegistry) and write too much code for being an “Architect”. I look back at nearly 30 years in professional software development and have seen the same fashion come and go a few times.

about the speaker:

I am Lead Architect in Microsoft’s Azure Messaging team that builds and operates a fleet of hyper-scale messaging services, including Event Grid, Service Bus, Event Hubs, Stream Analytics and Microsoft Fabric Eventstreams. I represent Microsoft in messaging standardisation in OASIS (AMQP, MQTT) and CNCF (CloudEvents, xRegistry) and write too much code for being an “Architect”. I look back at nearly 30 years in professional software development and have seen the same fashion come and go a few times.

description:

Mat and the team at Grafana Labs have recently launched an integrated agentic chat experience directly in Grafana’s frontend. In this talk, Mat will explore what it takes to build a truly useful Assistant—covering context and prompt engineering, effective tool design, flexible architectures, and more. He’ll also share details on their work to orchestrate a swarm of agents for observability, helping users quickly understand what’s happening in their systems. Expect practical, technical lessons from working hands-on with this emerging and unconventional technology.

about the speaker: 

I am an engineer at Grafana Labs, where I helped launch Incident, Drilldown apps, and the LLM-powered Grafana Assistant. A strong open-source advocate, I created Go’s most imported package, Testify. I am hosted the Go Time and Big Tent podcasts, and appears on Grafana Campfire calls. I enjoy solving problems with tech and has had multiple startup exits, including Machine Box in 2019.

🔒 topic locked

about the speaker:

Frontend developer in love with the Web, browsers, bicycles, and podcasting. He/him, Google Developer Expert. Based in Berlin.

description:

Writing documentation rarely tops anyone’s to-do list. It can feel slow and unnecessary—right up until poor communication, unclear goals or mismatched expectations turn into weeks, months and in some cases years of rework, frustration and endless meetings only to then culminate in an entire project getting tossed because it never addressed the real need.

This talk is about how one deceptively simple tool, the design document, can save your ML project from that fate. We’ll cover how it turns fuzzy ideas into concrete plans, exposes problems before they snowball, and keeps everyone aligned as things evolve. Along the way, you’ll see how starting with a few “boring” pages can be the difference between a smooth delivery and a costly post-mortem.

about the speaker:

I’m a Senior Applied Scientist at Wolt, where I work in our Logistics team to make sure you get your food as fast and as affordably as possible.

In my (still relatively short) career, I’ve checked off pretty much the whole corporate data science starter pack — churn models, ranking, time-series forecasting, image classification, and more.

And if from all of this I had to pick one thing I wish the younger version of me knew in advance, it would be this: a ML/AI-related project’s success or failure is dependent first and foremost on proper communication between everyone involved – which is where the design doc comes in 🙂

description:

In a Kubernetes galaxy not so far away, operators face a crucial choice for managing and securing their networks: the simplicity of an Ingress controller or the power of a Service Mesh. Service Meshes promise resilient deployments with automatic retries, TLS encryption, traffic management, observability, and more. But not all meshes are created equal.

From Istio’s scalability, Linkerd’s simplicity, and Cilium’s eBPF innovation to Kuma’s multi-cluster versatility and Ambassador’s edge connectivity, each offers unique strengths. With options like sidecars or sidecar-less setups, choosing the right solution becomes a battle of priorities.

Join Henrik from Is It Observable? as he unveils a benchmark comparing these Service Mesh contenders across critical dimensions: proxy type, feature set, user experience, observability, and performance. By the end, you’ll gain clarity on which Service Mesh aligns with your Kubernetes needs. May the Mesh be with you!

about the speaker:

I am a Cloud Native Advocate at Dynatrace and a CNCF Ambassador. Prior to joining Dynatrace, I worked for over 15 years as a Performance Engineer. I am also one of the organizers of the conferences WOPR and KCD Austria, and the owner of the YouTube channel Is It Observable.

we’ve reached the end of our program. let’s close the day with some final thoughts and quick recap. thank you for joining us at bit summit!

time to switch gears. join us for the after-party where we’ll trade whiteboards for cocktails and keep the conversations flowing. we tech hard, we party hard!