Big Techday 11

Adam Ondra talks not only about his successful climbing career but also about his life as such because climbing for him is a way of life. As he says, he is living the dream he had since he was a child. What does it take to get where he is? What is his key to his success? This and much more is the topic of Adam’s talk. He will also talk about his most recent successful projects such as climbing the legendary Dawn Wall, or climbing the world’s hardest route and establishing a new climbing grade 9c. All accompanied by stunning pictures and emotional videos.

0.18% or 0.43 points. After a thrilling competition this was the gap that granted our next guests' dream of winning an Olympic Gold Medal. While they were six points behind first place after the first part, the short program, they achieved a new world record in freestyle figure skating, which will be remembered for a long time to come. Our next speakers are figure skaters Aljona Savchenko and Bruno Massot, together with their coach Alexander König. This trio has been working together since the end of 2014 and seems to be unstoppable ever since. While they “only” scored a bronze medal at the World Cup 2016, they achieved a silver medal the following year and a gold medal at the World Cup after their gold medal at the Olympic Games. They did this with a world record score in the overall standings, as well as another world record for freestyle figure skating. In the following hour, we will get an insight into their time together in Oberstdorf as well as the ups and downs they experienced on that journey over the last four years.

What if augmented reality glasses could turn your environment into a huge spatial painting in real time, and everything would like an artwork of, say, Van Gogh? One of the many exciting developments in the field of deep learning is the so-called "style transfer". This is the possibility to create a kind of art patchwork (see pastiche) from two pictures. One image is the artistic style, while the other represents the actual content to be transformed. By using tools like OpenCV and TensorFlow, TNG’s Hardware Hacking Team has succeeded in creating such glasses. As a result, your everday environment can be displayed in the style of famous painters such as Claude Monet or Pablo Picasso. In this talk you will get an introduction to the world of real time style transfer. A special focus will be placed on deep learning techniques that can be applied to style transfer.

This talk gives an overview of cognition-enabled robot control, a computational model for controlling autonomous service robots to achieve home chore task intelligence. For the realization of task intelligence, this computational model puts forth three core principles, which essentially involve the combination of reactive behaviour specifications represented as semantically interpretable plans with inference mechanisms that enable flexible decision making. The representation of behaviour specifications as plans enables the robot to not only execute the behaviour specifications but also to reason about them and alter them during execution. I provide a description of a complete system for cognition-enabled robot control that implements the three core principles, demonstrating the feasibility of our approach.

Machine learning on Source Code is a new area of research in the field of artificial intelligence, which, unlike classical problems such as image segmentation, does not yet have established standard techniques. This technology offers a variety of possible applications, for example in the area of static code analysis or in the automatic selection of relevant test cases. In the talk, the colleagues of TNG Technology Consulting will present different methods for transferring classic machine learning approaches to this new field of expertise. For instance there exist standard methods for processing images, that make machine learning algorithms pay attention to their two-dimensionality. However, there are currently no common techniques for encoding the semantic structure of source code. Therefore, we had to find new ways to mathematically represent the code of projects and we will present our ideas to you in the first part of the presentation. Then you will learn about approaches for automatic as well as manual training data generation. Another part of the presentation includes an overview of different models and machine learning frameworks that seem suitable for this context. The last section deals with the possibilities of using such models for the analysis of code.

Neural networks outperform humans! At least in certain domains. Let's talk about when to apply neural networks and situations where other tools might be better suited. Recent advancements in hardware, software and neural network architecture enable us to create amazing products ranging from computer vision to medicine. Let's have a look at interesting prototypes from current research and the ideas behind them!

If you read the news today, you might come to the conclusion that the human race is doomed, that democracy is on the decline, that authoritarianism and tribalism are on the rise. But is the world really falling apart? In this bold keynote, Steven Pinker urges us to step back from the gory headlines, which play to our psychological biases. Instead, follow the data. Pinker shows that life, health, prosperity, safety, peace, knowledge, and happiness are on the rise, not just in the West, but worldwide. This progress is a gift of the Enlightenment: the conviction that reason and science can enhance human flourishing. But in 2018, have we become complacent? Have we taken the Enlightenment’s breakthroughs and ideals for granted? The Enlightenment stands against many of the darker currents in the air, which demagogues are all too willing to exploit, resulting in attacks on liberal democracy and global cooperation. In a timely and hopeful keynote, Steven Pinker, one of the world’s most influential public intellectuals, makes the case for reason, science, and humanism: the ideals we need to confront our problems and continue our progress.

Apache Nifi is one of the Apache Foundation's current open source projects with the goal to simplify application stacks of large companies using a highly versatile mediation system and to enhance performance in many ways. One can think of Apache Nifi as a data flow management platform. While the classic use case is ETL (Extract, Transform & Load), Apache Nifi can do much more. As such it offers a huge range of functions such as complex event processing, zero-master clustering, and connectivity to Apache Kafka, HDFS, MongoDB, Couchbase, HTTP, and of course also FTP. Especially by supporting historization of events and visualizations thereof (Provenance, Lineage), Apache Nifi offers more than most competitors. After an introduction to Apache Nifi and underlying concepts, performance and extensibility with own processors will be illustrated by using a case study from the field of telecommunications.

Rapidly increasing urbanization and our daily growing need for mobility are pushing our transportation systems and the global ecosystem more and more towards the limit. Clean and affordable on-demand air transport offers a way out by making use of the third dimension. For the first time in history, the overdue promise of flying taxis above our cities is about to become a reality. At the top of this revolution, Munich-based Lilium are working on the world’s first Electric Transition Jet. In his talk Co-Founder and Head of Autonomous Flight Control Matthias Meiner speaks about the future of a world where everyone can fly anywhere, anytime, and the challenges to be mastered along the way.

Robotics will substantially and sustainably impact our world, becoming one of the future central societal topics. Nowadays, however, it is only accessible to a limited number of people. High costs, complex programming and operation, as well as basic technological restrictions, which – among others – lead to absolute separation of robots and humans by safety fences, all represent impassable hurdles. Introduction of groundbreaking technology in robotics is now enabling a new generation of lightweight robots that will be able to conduct tasks that could never be tackled before.

Programming games has become much simpler in recent years due to the advent and easy access to professional gaming engines like Unity 3D, Unreal and Cry Engine. With these engines you also get access to a wide variety of code libraries, easy to use 3D models and more assets. This makes it possible to concentrate on the game concept itself instead of losing your focus in the details of 3D programming. Within this talk we will create an examplary simple Virtual Reality game in Unity 3D. The speakers will provide details about basic aspects of 3D software development like cameras, sky boxes, die illumination of the scene and the 3D objects as well as their components. Then they will give a short introduction into the program flow and the development of the gaming logic using a simple example. In addition to that, we will also bring up some issues concerning usability and maintainability.

The Robotic Operating System is a platform for the development of autonomous systems. Due to the flexible architecture and the consistent open source policy, ROS acquired a large supporter community right from the start. The platform, which was originally designed for robots, is now used for many application scenarios in industry and research, including autonomous driving. The new version 2 fixes many weaknesses of the first version. In this talk we give a brief introduction to the concepts behind ROS and its tools. Using examples, we demonstrate how the use of ROS can facilitate the integration of software and sensors.

Autonomous Driving Technology has the potential to fundamentally change mobility and how cities are built. However, this is a huge technical challenge, probably harder than putting a man on the moon! We believe that for overcoming this challenge it will be crucial to perfectly integrate insights from the software and the automotive industry. In this talk, we’ll review the state of the art in autonomous driving, some remaining key technical decisions (e.g. Lidar or no Lidar) and various paths to deployment at scale.

Today, we can fabricate anything. Digital fabrication now functions at both the micro and macro scales, combining multiple materials, and using different materialization processes. Complexity and customization are no longer impediments in design. While we can fabricate anything, design arguably appears confined by our tools of design: we can only design what we can directly represent. If one looks at 3D-printed artifacts, there is oftentimes a discrepancy between the wonder of technology, and the conventionalism of design. What is needed is a new type of design instrument. We need tools for search and exploration, rather than simply control and execution. These new tools should simultaneously be open and systematic, striking a balance between causality and chaos. They require a design language without the need for words and labels, as they should create the previously unseen. These tools must ultimately redefine the process of design: the designer will work in an iterative feedback loop with the machine, moderating processes, and incorporating feedback, surprises and proposals. Knowledge and experience are acquired through search, demanding heuristics that work in the absence of categorization. As of yet, we have countless tools to increase our efficiency and precision. Why not also create tools that serve as our muse, that inspire us and help us to be creative? Tools to draw the undrawable, and to imagine the unimaginable.

Martin Bellin is the founder and CEO of BELLIN. Founded in 1998 as a one-man company, he has shaped BELLIN into a key framer and the innovation driver of the fintech and finance industry. After studying business administration at the University of Mannheim, Martin Bellin worked for several years in the treasury of renowned companies. From his experience as a treasurer, he developed the vision of creating a technical solution for corporates that ensures the transparency and availability of group-wide financial data and facilitates global collaboration among the units. As none was available at that time, he built a Treasury Management System (TMS), from which the entire BELLIN range of offers emerged. His sense for the customers’ needs and his enthusiasm for perfection have repeatedly led to pioneering achievements and innovations. Martin Bellin has been awarded Entrepreneur of the Year several times.

Wolfgang Kalthoff is a member of the management board and CTO of BELLIN, being responsible for the product lifecycle - from product management and development to service & support. After completing his PhD in physics, Wolfgang Kalthoff began his career at SAP and initially dedicated himself to the development of software, software architecture, and project management. Later he moved to management and led international teams, i.a. he spent two years developing a local development department in Bangalore, India. Parallel to his professional activity, he completed the Executive MBA in Mannheim, then founded and managed a payment company. Since October 2017, he has been setting up the development teams at BELLIN international, promoting the technical rebuilding of the BELLIN platform and pushing ahead with new innovative products.

Classic blockchain technology provides certain security guarantees without implying trust on a third party. Nevertheless, all data is public. Still internal company data should remain secret and only accessible to legitimate interested parties. Private blockchains cannot fulfill many of their promises. A clever combination of classic enterprise databases with public blockchains can be used in an unprecedented way to protect data against manipulation. This talk gives a more detailed insight into the possibilities of modern public blockchains such as "Smart Contracts" and shows how these can be applied beneficially and cost effective in companies.

Once again, several million documents about tax havens and letterbox companies were leaked to the Süddeutsche Zeitung. The Paradise Papers document how criminal behavior is covered up and money from dubious sources is hidden. As with the Panama Papers, the SZ shared the data and again a global team of reporters worked to filter out the best stories. Working with such a wealth of data needs to be well organized and requires special measures to be taken. In this talk, Felix Ebert and Elisabeth Gamperl talk about the tools and technologies used to prepare, search, and visualize the data. In addition, they give an outlook on how to research in large amounts of data at the Süddeutsche Zeitung in future.

The Internet has been successful beyond even the most optimistic expectations. It permeates and intertwines with almost all aspects of our society and economy. The success of the Internet has created a dependency on communication as many of the processes underpinning the foundations of modern society would grind to a halt should communication become unavailable. However, much to our dismay, the current state of safety and availability of the Internet is far from commensurate given its importance. Although we cannot conclusively determine what the impact of a 1-day, or 1-week outage of Internet connectivity on our society would be, anecdotal evidence indicates that even short outages have a profound negative impact on society, businesses, and government. Unfortunately, the Internet has not been designed for high availability in the face of malicious actions by adversaries. Recent patches to improve Internet security and availability have been constrained by the current Internet architecture, business models, and legal aspects. Moreover, there are fundamental design decisions of the current Internet that inherently complicate secure operation. Given the diverse nature of constituents in today's Internet, another important challenge is how to scale authentication of entities (e.g., AS ownership for routing, name servers for DNS, or domains for TLS) to a global environment. Currently prevalent PKI models (monopoly and oligopoly) do not scale globally because mutually distrusting entities cannot agree on a single trust root, and because everyday users cannot evaluate the trustworthiness of each of the many root CAs in their browsers. To address these issues, we propose SCION, a next-generation Internet architecture that is secure, available, and offers privacy by design; that provides incentives for a transition to the new architecture; and that considers economic and policy issues at the design stage. Within roughly 100 person years, we have implemented SCION and deployed it in the production networks of two ISPs.

Over the past 3 years, ProtonMail has grown into the world’s largest encrypted email service, with over 4 million users. In this keynote I will discuss the benefits and challenges of working in the open and federated email space, the technical and usability hurdles surrounding encrypted email, and the importance of security-first development in the age of data breaches.

Statistical analyses of personal data have many applications: from medical research and traffic planning to the optimization of advertising strategies. Data protection regulations and rights are the major obstacles to such business applications. Consequently, in practice, businesses often try to persuade the affected individuals to renounce their rights in a more or less voluntary and usually very far-reaching way, e.g. imposing the acquiescence of dubious clauses in order to access online services. The legal hurdles to this approach have been intensified, effective as of May 25th of 2018, as part of the new EU General Data Protection Regulation. In this talk, I will illustrate that this matter is far beyond being unsolvable: technical automatization solutions with “Security by Design” allow us to perform statistical analyses of personal data without violating individuals’ privacy. An appropriate set of classic IT security mechanisms, modern cryptography procedures, and strict formal definitions of anonymity enables value-adding personal data analysis and guaranteeing at the same time a high degree of confidentiality, even in the case of a corrupted system operator.

Having small bundles of code is critical when you want your websites to load fast. So why not start by removing all these useless utility functions your dependencies sneaked into your code base that you neither use nor care about? With JavaScript's native module system, weeding out unused imports is usually referred to as "tree-shaking". But as it turns out, significant dead code elimination in a dynamic language like JavaScript is no easy task. To give you a glimpse of how this is actually done, I will show you how we reworked Rollup's tree-shaking algorithm to be able to track nested functions, return values, and object shapes, and how clean code principles helped us get there. And as a bonus, we will see how Rollup is now able to perform no-duplication code-splitting to utilise on-demand loading for optional code blocks using the host system's loading mechanism as the only runtime necessary.

DevOps teams share a very common challenge: Sharing passwords. While not encouraged, sharing passwords still is a common problem we needed to solve in the era of cloud- and service provider integrations. When crossed of cloud based and closed-source products for security and trust reasons, we decided to implement our own password manager using a git and gpg based storage format. We’ll show what gopass is, why it exists and how it might help you to solve some real world use cases. We further explain how gopass can be integrated in your workflow and why Go is such an awesome language.

Quantum computers and quantum simulators are computational devices the constituents of which are single atoms, ions, or other physical systems following quantum mechanical laws. The observation that a radically different physical theory compared to classical mechanics is at work might suggest different computational capabilities arising from such an architecture. In fact, such devices could solve some problems that are intractable on conventional supercomputers. Partially stimulated by a new impetus by research effort of US IT companies and the Flagship effort of the European Union, the idea of quantum computing and simulation - anyway only a bit more than two decades old - has gained much momentum recently. While quantum computers are so far still anticipated devices, quantum simulators exist to an extent. This talk will provide a brief introduction to the field of quantum technologies with respect to notions of quantum communication and in particular of quantum computation and simulation. I will argue in what way existing quantum simulators may already be seen as machines that (in a mild sense) outperform supercomputers.

Although the quantum computer is only expected in about 10-15 years, it is vital to take it into account already today. Actually, it is already too late to protect the confidentiality of information for several decades! Today, the only solution that is already commercially available is quantum cryptography. This technology may sound futuristic, but is actually already in use in some niche markets since about 10 years. Somehow, the theory of electromagnetism was discovered in the 19th century and was the basis for all information processing in the 20th century. Likewise, quantum theory was discovered in the 20th century and is likely to form the basis for the 21st century information processing.

When the paper "Bitcoin: A Peer-to-Peer Electronic Cash System" was published in October 2008 under the pseudonym Satoshi Nakamoto, the effects were hardly foreseeable. Ten years later, the topic of blockchains is on everyone's lips, especially in the form of Bitcoin. For this reason, this talk will deal with the functioning of blockchains, using Bitcoin as an example, and how, for the first time, independent of centralized trust instances, they will enable reliable transactions on the Internet. The central aspects include the cryptographic and mathematical basics, transactions, decentralized networks, as well as the authoritative "mining" based on the concept of the "proof of work". Finally, concrete applications of the blockchain are presented, i.a. in the areas of filehosting, publishing and copyright, social media, as well as insurance.

Laurent Jaffart is currently working in the Sales and Marketing organization of Airbus Defence and Space taking care of the future programmes with a focus on Space Systems. Laurent leads a trans-national team based in France, Germany and in the UK. The aim of his organization is to bring innovation on the market, being financial innovation, business model innovation or product innovation. Past assignments have taken him and his family across Europe and the USA. He started his career with EADS - in 2001 with a position in the post-merger integration team following the creation of the group after the tri-national merger of Aerospatiale-Matra, DaimlerAerospace (DASA) and CASA. He previously held a position with American Express EMEA Headquarter Project Office working on the introduction of the Euro. In 2002, he lead the EuroFighter Mission planner and Briefer programme to then take over in 2004 the MIDS LVT Programme - (Multifunctional Information Distribution System Low Volume Terminal Radio Frequency where he managed the industrialization of the programme into a serial production. From 2006 to 2008, he took a business development management position for the Defence Electronics business unit of the company, entering in the strategy arena of the company. From there, in 2008, he was assigned at the strategy department of the Defence and Security (ex- Cassidian) division headquarter to supervise and coordinate divisional strategic roadmaps, Merger and Acquisition cases and the initial cooperation with the former Astrium division of the group. In parallel, he led the programme management office for the transformation of the division establishing a matrix organization of countries and programmes. In 2011, he was promoted as Vice President of strategy integration leading the strategic planning process for the former Cassidian Division. Following a year at the National Defence University in Washington DC at the Industrial College of the Armed Forces (now Eisenhower School), he joined the Astrium division as the Vice President Head of Astrium Satellite Strategy in 2013. Since 2015 and while becoming Vice President Head of Future Programmes in the sales and marketing organization, he led the OneWeb bid for a constellation of 900 telecommunication satellites and the creation of the Joint Venture - Airbus OneWeb Satellites - to design and manufacture constellation satellites amongst other opportunities and most recently was appointed as Airbus Board Member for EarthNow a new venture delivering true real-time video of the Earth that is persistent through a large constellation and enhanced by machine intelligence. Laurent Jaffart holds a Masters in National Resources Strategy from the National Defense University (2013), two Masters in Business Administration (MBA) from the former ESC Reims (Reims Management School - France) and Dublin City University (2002) and a Bachelor of Sciences (Hons) in European Business and Technology from the University of Brighton in the UK (2000).

A proper architecture is essential for any major successful software product. However, the documentation, implementation and dissemination of knowledge is an ongoing challenge that seems to grow exponentially with the number of developers involved. How do we prevent our designed architecture from becoming a messy concept at the documentation cemetery? How do we create a common understanding of all involved developers? And how do we prevent the actual and the target architecture from completely diverging and ending in the famous Big Ball of Mud? In the first part, this talk deals with these questions in an abstract manner. In the second part I introduce the Java Open Source Library ArchUnit, which automates many features of code and code structures as simple unit tests and ensures them over the long term. From typical automated checks, such as "Package com.foo can not access package com.bar", to complex rules, such as "Classes annotated with @ClientEndpoint and located in the package my.app.client.v1, can only use methods whose return values are serializable” ArchUnit offers a great deal of flexibility and can be adapted and expanded in many ways to meet your own needs. Many hands-on coding examples that shed light on these possibilities will be given.

Most organizations underestimate the value of a mature "Run" organization which must be the foundation to successfully embrace innovation and to make the right choices. Any new project consideration should anticipate the "Run" impacts and will only be successful if the basics are under control.

Build systems are awesome - but unappreciated. A decade ago I started writing Shake, a Haskell library for implementing build systems, an alternative to Make, but with much more powerful and accurate dependencies. In this talk we trace the origins of Shake, from a small module to build a PhD thesis, to a production quality build system, figuring out the key ideas along the way. By exploring these ideas we see how build systems relate to each other, realise that Excel is really just a build system, and see how Bazel/Buck keep shared caches in the cloud.

As developers and architects, we spend most of our time further developing existing systems (i.e., “maintenance”). Therein, we are often annoyed by the many deficits of these systems: Even supposedly simple things become unspeakably difficult when dealing with _legacy systems_, the _time-to-market_ gets worse and worse, and the departments build up more and more feature delivery pressure. In this talk, I will show you ways to systematically escape this trap that is caused by technical and general deficits: Get to know long-term improvement approaches, with which you can solve even the most muddled situations. For that purpose, we first go on a methodological breadth-first search for problems in order to clearly identify our technical, organizational and communicative deficits. Subsequently, I will introduce you to a few strategic improvement approaches, such as:
* Reduction: migration towards self-contained systems ("change-by-split" and "change-by-extraction")
* stronger focus on the subject matter: a step-by-step introduction of domain-driven design into legacy systems ("restructure-to-domain")
* Enhancement of cohesion within systems ("improve modularization")
* Homogenization – solve similar problems in systems with similar ways ("improve consistency")
* Clean-Up of data: Modularization and homogenization in inventory data ("database refactoring")
I distinguish these strategic (i.e., long term) approaches from the known tactical improvements, such as refactoring, and illustrate them with the help of (anonymized) practical examples.

Don't grow up, it's a trap! On April 14th, 2018, seven teams consisting of a total of 27 people set out to build the LEGO Millennium Falcon (set number 75192) within a mind-blowing time of 2 hours and 10 minutes. In a massively parallelized setup, the teams provided the sub-assemblies that were then used to assemble the model. In this talk, we will dive into the methodology used to organize such an ambitious project. Furthermore, we will investigate what we learned from our first attempt, which resulted in a 2-hour 47-minute record, and how we managed to improve our time by roughly half an hour. Additionally, we will show interesting parts of our livestream video that highlight key parts of our preparation and methodology.

This talk should give you some notes on the investor’s/VC’s view of tech companies. After a brief overview of the market situation, we will particularly discuss the relationship between investors and technical organizations/CTOs. In addition, we go through some expectations and requirements in the different phases (early stage vs. later stage). In the end you should have a better understanding of the investor‘s mindset and his way of thinking. Of course, there will also be enough time for questions.

In addition to strength, endurance and technique, above all the mind decides about victory and defeat. Consistent work on the right mindset has now become a matter of course, so that it pays off for anyone, who wants to get their “horsepower” focused on the road in everyday work without being left behind, to think outside the box. In this talk, know-how, concrete methods, and tips are conveyed to illustrate how to reach high performance, without risking physical and mental health. Without esotericism and incense sticks, well-founded techniques for task and mind management as well as visualization and focusing are at the core of this talk. The better people understand how they function, the better they can build a mindset where career success and private satisfaction are not opposed but in balance.

Autonomous driving solutions introduce a new complexity into the development of embedded systems in a car. This complexity rises with each level of control and autonomy. New tool categories have to be added like machine learning, but also existing technologies, like simulation, are stretched to their current limits. The toolchain for such challenges is complex and the integration of all the tools coming from different domains cost a lot of effort for each market participant without a real competitive advantage towards the solution. Therefore, the Bosch Autonomous Driving division, together with many partners, is currently building up the OpenADx ecosystem including OEMs, tier 1 suppliers, tool vendors, research organizations but also companies from related industries like the IT industry. The goal of this endeavor is to stop wasting money on the introduction of a proprietary toolchain in each of the companies and to share the development costs. Besides the benefit of sharing the costs, the expected result is a better integration of the toolchain within the organizations, but especially also at the interface between cooperating organizations. For tool vendors and research organizations, the advantage of the approach is in the existence of an integration backbone, which allows the provider to easily integrate new technologies or tools into a working environment that runs in a multitude of customer organizations, instead of providing proprietary solutions for single customers. This does not mean to replace existing tools. There are many tools and technologies, be they commercial or open source that solve perfectly their job and should do so in the future. The initiative targets at an optimal support of the development workflows along with the integration of the relevant tools as well as filling in gaps identified and not solved by existing tooling. This integration glue will be provided as open source software under the umbrella of the Eclipse Foundation. The current state of the initiative is that there is a substantial amount of interested partners which work together to concentrate on clearly scoped topics that are of general interest. E.g., some partners are working on a demonstrator that combines several tools in the area of simulation using DDS as a communication technology. A second topic identified is the management and ingest of massive amounts of data. Here the current challenge are the creation of importer for standard data formats and building up a corresponding infrastructure. A core technology identified is ROS, especially ROS2. As part of OpenADx, Bosch supports the porting of functionality like RVIZ and Rosbag from ROS1 to ROS2. OpenADx is an industry approach to collaborate in non-competitive areas to set up toolchains with less effort to accelerate the development of Autonomous Driving solutions.

Why is architecture necessary? When automating business processes, it is necessary to look at them holistically, align them to the company's long-term strategy and shape the future. The architecture department specifies framework conditions for the development and uses a broad toolbox of communication, documentation and planning methods. When creating solutions, different approaches come together constructively to achieve the best result for the company.

"Time to Market", innovation management and agility are terms that are wide spread today in almost all technology companies. The results are usually more modest than the expectations and it is not uncommon that they remain buzzwords of Powerpoint presentations. We started working on these areas about a year ago and are already seeing significant (and measurable) successes. Stefan Seuferling will give an overview of the activities at Dräger and some impulses.

• Setting the right goals / strategic goals as guardrails

• Overall concept and overview of the change roadmap

• Accompanying initiatives & what really mattered to Dräger

• Tracking and motivation

• Right people on the bus (role models, competence management)

• Introduction of agile leadership

• Next level

In many (software development) projects, cost estimations - often in abstract points - are used as a basis for decision-making. However, it is rarely examined how reliable this basis really is. How high is the risk that the effort will be significantly higher or lower than estimated? Does an estimation of three points today mean the same as one half a year ago? And is the scale used even appropriate for the achievable accuracy? Based on data from estimates and actual costs gathered within two years in a major TNG project, I investigated these questions with statistical methods and derived recommendations for dealing with cost estimations.