Programme 2021





Admission and registration

Our location: OSTRA-DOME Dresden


Opening Speech: Frank Wolf, Founder of Staffbase


Stubborn on the vision – flexible on the details: How Amazon teams build for speed and stability

Markus Ziller, Senior Solutions Architect @ Amazon Web Services

Organizations of all sizes often ask how they can innovate like Amazon. Since its humble beginnings as the "world's largest bookstore," Amazon has not only innovated across ecommerce and other adjacent segments, it has also introduced entirely new and unrelated businesses to the organization. Over the years, Amazon has gotten good at taking problems that are hard to solve and making them easy for customers. In this session, we discuss how Amazon views and leverages technology as a utility to increase the pace of innovation across the organization.

From TensorFlow to beyond: how to use a TensorFlow-trained model in the cloud, mobile, and web

Juan De Dios Santos Rivera, Data Engineer @ Lovoo

There’s more to TensorFlow than fit and predict. It is a vast ecosystem that covers multiple platforms such as mobile, cloud, and web. In this talk, I’ll present how you can train a model in TensorFlow and export it to a mobile device (using TensorFlow Lite) to the cloud (using Google Cloud’s AI Prediction) and to the web (using TensorFlow.js). Moreover, I’ll describe TensorFlow’s main export format, named SavedModel, and small tips I’ve learned during my adventures across TensorFlow Land.


What's going on under the hood? Getting concurrency right in Arm processors

Diogo Behrens, Research Engineer @ Huawei Dresden Research Center
Nowadays, Arm processors are everywhere (mobile devices, but also more and more in the cloud as servers, example Amazon Graviton EC2). If you work on system software or legacy software (Linux, MariaDB, etc) and you need to run that on such Arm processors, you’ll have a lot of issues with concurrency because of the memory optimizations of these processors. Diogo focuses on one class of algorithms in the presentation: synchronization primitives such as mutexes and spinlocks and he explains what is the actual problem with these memory optimizations, why it causes the software to crash or hang if you are not careful. And finally will talk about the results he has.

Organizing App Teams for fast Flow and Autonomy leads to Product Development Nirvana

Ralf Jacob, Head of Technology – Mobile @ idealo internet GmbH

A practical example of how to become independent within a larger tech organisation. A Self Contained Systems architectural approach allowed a small product team to deliver business value with native applications at unrivaled pace and with drastically improved happiness.



Rewrite or Refactor That Is the Question

Vladimir Jovanović, Senior Android Engineer @ Flixbus

Your app is in really bad shape, and you are not sure if you should refactor it or go for a rewrite. This talk will give you all the information necessary to make the right decision. We will follow the Android and iOS teams at FlixBus fighting with legacy code, each in their own way. The Android team chose to refactor their old app, while the iOS team decided to create a brand new application.

Automatic Data Extraction in arbitrary Documents

Bianca Preißler and Toni Das, Machine Learning Engineers @ elevait

We will talk about how to bring automatization to an industrial context, using machine learning, especially computer vision. Exemplarily, as a showcase, we will present a domain-agnostic workflow to extract information from documents and discuss techniques, challenges and successes along the way.



Breakfast & Networking


Pipelines Are Awesome: SAP’s Journey to a Scalable Mobile CI/CD

Philipp Herzig, Head of Product Engineering Intelligent Enterprise & Cross Architecture and Marc Bormeth, DevOps Engineer @ SAP SE
Building stable and scalable continuous integration and delivery (CI/CD) workflows gets more complex with a growing number of teams and projects involved. Join Philipp Herzig and Marc Bormeth for this session to learn how you can overcome common challenges. Effectively composing publicly available tooling like fastlane, piper, and Azure DevOps is key to make developers’ life a little easier. If you are curious how to drastically reduce the setup time for new projects, improve reusability and scalability, and see hands-on examples from SAP, this talk is for you. Let's focus on what is really important: Building great mobile apps!

Intensive workload management on decentralized cloud infrastructure with Krake

Hannes Braun, R&D Software Engineer @ Cloud&Heat Technologies GmbH

We will present the capabilities of the open source software Krake on how to manage the lifecycle of computationally intensive workloads and their distribution to the decentralised infrastructure, e.g. spawning new jobs or increasing and decreasing the number of working containers with special attention to optimised ecological energy use. In addition, the functionalities will be briefly deepened in a short demo during the presentation.


Defend your software from yourself

Maksim Gudow, Senior Developer @ DevBoost GmbH

Software developers are humans. And human beings are vulnerable to all kinds of biases. I would like to show you the tricky ones and give you advice on how to avoid them, making your software better in the end.

VIRTUAL: Understanding Azure API Management - 1

Sven Malvik, Team Lead Cloud Platform
@ Vipps AS

APIs are the front door to our micro services. Sometimes we let requests just walk through, other times we do a full body check and give them some additional information on their way. This hands-on tutorial will improve your understanding of Azure API Management.


Applying Innovative Technologies in the Enterprise World

Adrian Kostrz, Innovation Manager AI & Robotics & Matthias Siegmund, Innovation Manager @ NTT DATA Business Solutions

Technology for technology's sake is meaningless. Innovative technologies have made it possible for companies to transform their business models and to better interact with customers. However, as an increasing number of businesses jump on the bandwagon for the latest technology, the question arises whether people are driving technology or are being driven by it. Within this session, we will demonstrate different innovation cases and explain the usage of drones, image recognition, big data, and artificial intelligence in an enterprise environment.

VIRTUAL: Understanding Azure API Management - 2

Sven Malvik, Team Lead Cloud Platform
@ Vipps AS

Double session, second part.


Lunch & Networking


Meetup Expo


Moving real-time communication services to the Cloud with the power of Kubernetes

Sebastian Daehne and Alfred Krohmer, Sr. Manager DevOps & Sr. Software Engineer @ LogMeIn Germany GmbH

Realtime applications do have different operational requirements than for example micro services. We would like to show how we were able to create an environment for such services in the public cloud by using Kubernetes/Container primitives.

Machine learning at the IoT edge

Florian Mair, Solutions Architect @ Amazon Web Services

Whether it's connected cars, smart home devices or industrial solutions, IoT applications are rapidly becoming smarter. Edge computing is helping to drive this transformation. IoT devices not only collect and transmit the data, but also perform predictive analytics and respond to local events, even without cloud connectivity. Learn more about ML Inference at the Edge, why it matters and how you can use it to create intelligent IoT applications.


What’s currently hot in AI – review of the latest research findings

Torsten Hartmann, CEO @ Avantgarde Labs GmbH

Researchers and practitioners have pushed the state-of-the-art in many subfields of artificial intelligence in the last years. The presentation gives an overview over some of the most promising publications from leading AI conferences like ICML, NIPS or ACL and leading corporate labs like Google Research or Baidu Research.

Myths and Benefits of 5G and Beyond?

Vincent Latzko, PhD Researcher @ TU Dresden

5G has been heralded as the enabler of the next industrial revolution, a prerequisite for autonomous driving, and everything inbetween ('tis not a talk about covid-rays). Now that we have it, we wonder where the promises went: did they smoothly flee to 6G? What exactly was promised, what subset of that was realistic, and what is the technical basis for these visions? Can we still achieve Great Things without waiting another decade (the next G), can or should we put a focus on dataflow privacy, and how is AI helping?


GitOps using Flux: How we manage Kubernetes clusters at Staffbase

Rico Berger, Site Reliability Engineer @ Staffbase GmbH
At Staffbase we are using GitOps, to implement Continuous Deployment for cloud native applications. It focuses on a developer-centric experience when operationg infrastructure, by using tools developers are already familiar with. The core idea of GitOps is having a Git repository that always contains declarative descriptions of the infrastructure currently desired in the production environment.

VIRTUAL: Evolution 3.0: solve your everyday problems with genetic algorithm

Mey Beisaron, Backend Developer @ AppsFlyer Inc.
When was the last time you wrote an algorithm to plan your diet? In this talk I will share how I coded a Genetic Algorithm and used it to solve my everyday problems: planning my diet and semester timetable. Using code examples I will bring you closer to solving your everyday problems with GAs.


Coffee Break & Networking


Learning Machines - Machine Learning and Industrial Prognostics

Steffen Seitz, PhD Researcher @ TU Dresden
Monitoring the condition of machines is one of the most important use cases for machine learning in predictive maintenance. In order to train respective models, modern industrial facilities produce terabytes of unlabeled data every day. Unfortunatly, supervised learning approaches rely on labeled data to perform. In this talk, we will discuss this dilemma of supervised learning in industrial prognostics and show how to train meaningful models in order to detect machine failures earlier than ever before.

Building a Central Nervous System for Data with Apache Kafka

Anatoly Zelenin, Freelancing Kafka Expert and Author

Do you want super-powers? Analyzing and reasoning about data in large companies in real-time is one of these! Apache Kafka is not only about moving data between services, but about building a backbone for your data infrastructure.
With such a backbone you know what happens in your company NOW. In this talk, we will explore how Apache Kafka can help you to achieve this and cover how you and your company can avoid common pitfalls. Get ready for the wonderful world of real-time data!


You don’t need your freaking Kafka - Event Bus for small teams

Ilja Bauer, CTO & Co-Founder @ Entwicklerheld

Many service architectures use an event bus to integrate multiple services together. The use of an event bus offers many advantages in the way services can be developed, tested and operated. The use of common event bus technologies such as Kafka, RabbitMQ or NATS is associated with a steep learning curve and is not feasible in operation, especially for small teams. However, most teams have a lot of experience with classic relational databases like PostgreSQL or MariaDB. In the presentation, we explain how we use PostgreSQL & Redis as an event bus for over 1.5 million events and thus use the advantages of an event-driven integration for our EntwicklerHeld platform right from the start, without obstruct the switch to a "real" event bus later on.

On the necessity to fail from the perspective of a company that must not fail

Andreas Nicklas, Chief Digital Officer @ SachsenEnergie

As an energy supplier, SachsenEnergie AG operates critical infrastructure. Security and stability naturally come first. In the competition for customers, we also have to be innovative and agile. A field report on squaring the circle.


Digital Twin: Cloning real-world robotic devices into the cloud

Florian Purchess, CTO @ WAKU Robotics

A robot revolution is happening right now, that dramatically impacts the way our logisitics and production facilities work. Autonomous mobile robots are springing up like npm packages, covering our shop floors and warehouses. But new technologies always come with their own unique challenges. This is a tale on how we´re tackling them at WAKU Robotics.

VIRTUAL: Technical SEO demystified for web developers

Martin Splitt, Developer Advocate @ Google
The web is a great platform to reach a broad audience with our web apps and sites, but to do so we need to make sure that our web apps are findable for our potential users. In this session we will open up the black box of SEO and find out what we can do in our code to make sure our web apps are search-friendly and gather a set of tools to debug and test our sites for discoverability.


Closing Talk


Dinner & Networking