Dev@CGI Congress

How Low-Code Can Fuel Innovation for Pro-Code Developers
Alexander de Kriek

Innovation can be accelerated by faster Development Cycles allowing to build and test new ideas which allows experimentation and Rapid prototyping.

Especially with Low-Code the focus can be on the Core competences to accelerate innovative solutions and improving user experiences.

Most of the Low-Code platforms offer Collaboration tools to involving multiple teams in the development process and help to bring products to market more quickly

Low-Code platforms empower developers to focus on the creative aspects of software development, accelerating innovation and driving business growth.

The secret sauce of iAMLAB
Henk van Haaster/Rens van der Meijs

At iAMLAB, we develop innovations with impact and offer guidance in operationalizing proven techniques. What makes iAMLAB stand out Is our unique approach, our multidisciplinary team and active client involvement. So how does it all fit together? In this presentation, Henk and Rens will demonstrate what innovation looks like at iAMLAB as a Software Engineer. Find out how we create value for clients and partners using hackathons and how we continuously evaluate our topics at the iAMLAB for their relevance and feasibility.

OpenAPI – definition first
Wouter van der Harg

How to define and implement an OpenAPI serverless application. Generate the gateway, validate the input and store in a NoSQL database. All of this specification first, instead of code first.

My adventures as a woman in IT
Marion Braams

My adventures as a woman in IT since the 80 of last century. When I started most IT people were men. I started as a backoffice in Shell and then became a developer. Actually, my biggest project was building EuroShell, you know, the fuel pass we use every week.

Building deep learning apps for coders
Miguel Faase

Many great ideas start from hacks and gadgets. Given the abundance of open-source tooling, hacking AI-powered apps together is easier than ever.

In this live coding session we will learn how to  use the Huggingface ecosystem to access open source models for your own projects and host them on your own systems. The talk briefly introduces principles of deep learning and the types of tasks deep learning models are typically used for.  We will explore the models available and select a few from the ~700.000 models available to us right now.  The transformers library offers a convenient API to access the models. We will use this library to quickly develop an application that covers multiple deep learning domains (vision, language, audio), package it, and deploy it as a service. If time permits us to do so, we can attempt to retrain a model on our own dataset.

Attendees are invited to code along if they want to, but we will not have the time to wait for everyone to get set up.

• Conda package manager
• Docker engine
• Your favorite editor

From Normalized Systems Theory to Rejuvenation Factories
Herwig Mannaert

Back in 1974, Manny Lehman stated in his law of 'Increasing Complexity' that the complexity of a software system increases unless work is done to maintain or reduce it. Fifty years later, maintenance and evolution of software systems are still seen as problematic, often leading to decommissioning, and remaking the software.

Normalized Systems Theory (NST) provides a theoretical basis to explain and mitigate Lehman's law. According to NST, one change somewhere in the code often leads to other changes that propagate through the structure. Without proper and active design and maintenance, this is a major cause of increase in complexity.

NST proposes four design theorems that software designers need to adhere to, and five design patterns that enable developers to mitigate these ripple effects.

Based on these principles, our company, NSX, has developed code generators, called NS Expanders to automatically generate and rejuvenate at scale software systems that weakens the ripple changes and mitigate the increase of complexity.