• About DIVISIO
  • AI Basics
  • Amazon DJL - a new DL framework for Java
  • BGL symposium 2019 - lecture 'AI and Magic'
  • Categories
  • ChatGPT
  • ChatGPT for Teams: Privacy-Compliant Use in the Workplace
  • Conferences
  • Contact
  • Deep Fakes - How to spot faked Images
  • Deep Java Learning Introduction - Part 1: NDManager & NDArray
  • Depixelizer
  • DIVISIO Blog | Understanding AI, DL and ML together
  • DJL
  • DL4J Workshop at the ML Summit in Berlin
  • Enterprise Tensorflow - Java vs. Python
  • Enterprise TensorFlow 2 - Saving a trained model
  • Enterprise TensorFlow 3 - Loading a SavedModel in Java
  • Enterprise TensorFlow 4 - Executing a TensorFlow Session in Java
  • Enterprise Tensorflow: Code Examples
  • Git as a management tool for training data and experiments in ML
  • GPT-3
  • GPT-4
  • Homepage
  • Imprint
  • Java
  • Jax 2018 - Talks about DL4J and more
  • Jax 2019 Recap
  • KI Basics
  • Konferenzen
  • ML Conference 2017 in Berlin
  • MLOps: Establishment and operation of an AI
  • Neural networks - The five most common mistakes
  • NLP
  • NLP, NLU and NLG: AI and text
  • OpenAI API
  • Privacy Policy
  • Recap: ML Conference 2019 in Munich
  • Sokratext NER - Analyze German texts using Java
  • Sokratext NLG - Generate thousands of German texts with just one click
  • Sokratext POS Tagger - Automated Word Type Recognition with Java
  • Sokratexts
  • Tags
  • TensorFlow
  • TensorFlow and Java - An interview with entwickler.de
  • Types of Artificial Neural Networks
  • Understanding AI - Part 1: What is AI?
  • Understanding AI - Part 2: Symbolic AI, Neural Networks and Deep Learning
  • Understanding AI - Part 3: Methods of symbolic AI
  • Understanding AI - Part 4: The basics of Machine Learning
  • Understanding AI - Part 5: Supervised & Unsupervised Learning in ML
  • What are Neural Networks and how do they work?
  • Whisper 3 Large for JAVA
de en

About DIVISIO

DIVISIO offers you consulting and development in the field of machine learning and artificial intelligence. Whether as a finished product or as an individual customised solution - we successfully integrate AI to your company. Contact us at sales@divis.io.

Our services in detail

As a customer of DIVISIO, you will have a committed and helpful solution provider - We accompany you from brainstorming and planning to implementation and integration of your AI software. We are also happy to maintain, service and improve your finished solution within a maintenance contract. Depending on your requirements we will take care of all development tasks or work in cooperation with your team. Before the start of your project we are happy to provide information and training on the topics of artificial intelligence (AI), machine learning (ML) and deep learning (DL) through lectures and workshops at your company We can provide our services for all phases of a typical AI project:

Kick-Off and brainstorming

  1. Consulting and analysis Together we look at how we can profitably introduce AI and machine learning to your company. You will learn how a successful AI project works and which steps are necessary for your specific goals.
  2. Brainstorming In this phase, we define measures to achieve the goals we have set together. We determine exactly which process makes sense for your application and what data is needed for it.
  3. Software architecture We ensure that the AI system to be created can be robustly integrated into your enterprise environment.

Data acquisition and processing

  1. Data acquisition Together, we establish an overview of the available data, how to consolidate it and how to persist it in the right format.
  2. Data analysis This step will make sure all data is used correctly and the resulting system will be trained in the most efficient manner.
  3. Data processing In order for the data to be of use for machine learning, it needs to be cleaned and preprocessed based on the findings of the analysis step.
  4. Preparation of test plans and quality control Once all the data has been collected and processed, we ensure its functionality and quality by establishing test plans and quality control schemes.

Implementation

Implementation of your AI system can be done with one of the following approaches, depending on your needs:

  • Development and training from scratch We develop a unique software from scratch for you and train this model.
  • Based on a DIVISIO product, followed by training / finetune An existing DIVISIO product is trained and adapted to your requirements.
  • Based on a third-party solution We integrate a third-party product (Alexa, Watson, Google, etc.) or open source software and adapt it to your use case.

→ Integration into your IT system With all three options, the resulting AI will be integrated seamlessly into your IT infrastructure.

Deployment, further support and maintenance

  1. Choosing the right hosting variant Your custom-made AI solution can be hosted on-site on your own infrastructure, with third-party cloud providers or in our data center.
  2. Support and maintenance Tuning and retraining of your model, support and maintenance as well as troubleshooting are ongoing services we provide for you.
  3. Further training We organise and accompany training courses for your employees on how to use and maintain the finished system.

Kick-off and brainstorming

The first step necessary to provide the best service to your business right from the start is a precise requirements analysis. This is an important first step to determine the right use of ML technology at scale and to best benefit your company. Our experts will be happy to advise you on the possibilities of modern machine learning algorithms and proven AI processes. To ensure that the desired results can be achieved at a high quality, we will make it our job to understand your business model and your in-house workflows. For you, this means a resource-saving gain in time, money and quality. Based on the requirements, we carry out a feasibility analysis using prototypes. This allows us to make adjustments to the goal or scope of the project at an early stage and save additional integration effort later.

Regular adjustments at an early stage

Especially in the beginning of a project, data sets are continuously adjusted and reevaluated. By trying out several possibilities, we finally determine which algorithm delivers the best performance for your task. Once a commercially viable use for your company has been found and its feasibility confirmed, we plan the interaction of a finished AI component with your company’s existing IT systems. Except for start-ups, integration into proven and business-critical enterprise environments is usually necessary. Based on our decades of experience with commercial enterprise systems, we create an integration concept for you. To ensure a smooth integration of your AI, we work in close cooperation with your team.

Data acquisition and processing

Conventional software is determined by source code written by humans and the input of structured data, such as in databases. With machine learning technologies data takes on a new and even more important role: The behaviour of the software is learned from existing data. Data thus takes on the same significance as the source code that was previously created by hand. Therefore, it is absolutely essential for the success of your project to have a lot of clean data with which an AI model can be trained. Depending on your requirements, we help you choose the right software to annotate and prepare your data or even develop custom software if necessary. Once all the data has been correctly collected, imported and evaluated, we sift through it in detail. Using visualisations and statistical analyses, we determine the necessary processing steps that lead to the best possible trained AI system for your requirements. Finally, we define quality criteria to ensure performance during development and operation of the finished system. For this purpose, we create suitable criteria with corresponding test procedures according to your requirements. We integrate these into the automated pre-processing and training process of the AI system.

Implementation

The type of implementation is chosen individually based on your requirements and already existing solutions in your IT infrastructure. Depending on the technical and business requirements we either develop a completely customized software, adapt an already existing DIVISIO product, adopt an existing OpenSource model or integrate the AI API of a third party supplier like Google, Amazon, IBM or DeepL. The questions we ask ourselves in this process are:

  • Is this a unique problem?
  • Can an existing solution be used?
  • Do certain data protection requirements have to be met when collecting and processing data?
  • Can a third-party provider be used or must the data not leave Germany (or even your IT system)?
  • On which systems should the software be operated?
  • Are there existing requirements for the selection of programming environments and libraries?

Individual software and DIVISIO products are usually implemented in Java - this makes them stable and flexible for most enterprise requirements. We also offer implementation in Python (e.g. with TensorFlow, Keras or PyTorch) or for .net environments as well as hybrid solutions of TensorFlow and Java. How we integrate our software with your systems depends on the chosen technical implementation. The AI system can be integrated as a library into already existing systems in your company, or it can be implemented as a separate Microservice.

Deployment, further support and maintenance

At the end of the implementation you will have a customised AI system that can be used profitably based on your data and business know-how. To complete the process together with you, installation (deployment) in a data centre or with a cloud provider is required. We offer the following options for installation:

  • on site at your company
  • on site with us
  • with a cloud provider
  • with another service provider

After completion of the project, we offer ongoing support, bug fixing and monitoring to ensure constant quality. In many application scenarios for machine learning, new data accumulates after completion of the first version, which can be used to improve the performance of the system. Through constant training and updates, we will continue to increase the performance of your AI solution. To protect your investment in this new technology, we also offer training on the new system for your employees.