New technologies are often used as a buzzword, for us they are importants building blocks in creating a new target architecture that can benefit the industry.
We don’t see Machine Learning as a universal brain but as a set of algorithms that can be used to generate heuristics to solve a complex problem with small steps.
Machine learning has proved to be very efficient in learning from data in order to recognise pattern shape one of the most successful applications has been the automated translation as done by google translate. The training of such engines is done by feeding the system with books already translated in different languages.
Data models for computer systems can be assimilated to language for humans, they are different ways to express a common fact. But they are more structured and simple than human language.
Bailam has applied the core concept of machine learning of using data to understand logic to data models. Most ML models used today are designed to learn with large amounts of unstructured data and to produce classification forecasts or trends. Our need is different as we need to produce an exact relation between data and a small sample of data. Bailam tram with the help of Machine Learning experts has developed a unique proprietary methodology to solve such cases.
The results have been very positive with a high level of accuracy with less than 100 records.
Application Programing Interface architecture , Open API
After the age of monolithic system and object programming the enterprise has embraced the API and service architecture that allow different technical components to communicate even if they are using different programming language, different platform and physically on different servers.
The technical issue of safely connecting these components has been mostly solved and secured REST API are commonly used.
In order to help the developer in the use of these API a Open API standard has been developed for the description of these interfaces. These documentations are very useful for technical API that demand little business understanding but are not enough for exchange of data that contain complex business concepts. Open APi has been developed by developer for developer and is hard to use by business analysts.
Bailam approaches leverage the OPEN API approach by using the examples available in the documentation to provide a business understanding of the data and to automate the mapping.
We believe in the strength of collaboration inside teams, across teams and across enterprise. The cost involved with integration does not benefit anyone and is only an expensive pain for each party.
Software vendor would benefit of creating community among their clients.
Bailam will empower the community to share securely the description of business scenarios that become the base of the process for integration. These communities could be global internal to a company or shared between clients of a software.
DevOps power of automated testing
DevOps has been a revolution for the cloud application and the important investments to automate the deployment and testing has paid off and today nobody would think of going back to manual process.
Today's full DevOps circle is hard in the enterprise world because business software packages are not ready for it yet and because the testing should include end to end business scenarios instead of step by step technical tests.
With standardisation of business test scenario and automat mapping Bailam brings a missing piece to the future of DevOps in enterprises.
Beyong No Code development :
No Text Specification
Over the last 10 years the No Code solutions have taken more place in the development of applications. No Coe solution allows the development of application via graphical interface to either draw interface or draw processes that are translated into code without asking the developpeur to see or understand the code.
These tools simplify the developments once the developer knows what he wants to develop. In the situation of integration or migration the developer or BA doesn’t know what has to be developed and the specifications were clear the coding would generally not be hard. But in most cases the documents that describe the source or target data are not up to date and the only way is to analyse the data directly.
Bailam approach facilitates the job of BA and developer by doing the analysis of the target and source automatically and by suggesting logic that could solve the problem. Instead of having to read an uninviting data dictionary and to write a painful specification document the team has to share sample data and Bailam deduct the logic from the data. This approach makes the project faster and improves the communication between the teams.
Programming by example
If you try to explain algebra to your kid you quickly understand how learning by examples works. The selection of the examples is important and allows the kid to find how to differentiate the different cases and similarities. For Bailam it is the same: the selection of scenarios used for the learning is important so the data is not seen as an enemy but as a friend.
At the heart of Bailam engine is the concept of programming by example. This concept has been around for many years, expert systems were an example of it. They have generally been tried to resolve complex logic and inference models. Bailam approach focuses on using programming by example to automate a large number of simple rules which is usually the situation in enterprise integration and migration.