We developed an entire machine learning platform and helped aggregate data from all the existing systems.Current status is: IN PRODUCTIONThe ML Platform starts at the data scientists onboarding and ends in production. No need for rewrites. A framework was developed to help the data scientists produce production ready deliverables while allowing them to explore the data. The Cloudera cluster is leveraged to aggregate the data from differents live and legacy systems. The Kubernetes cluster and cloud infrastructure are leveraged to allow serving the models in a traditionnal way. The CI/CD helped package the Python into reproductibles build. The data was versionned and linked to the deployments.The stack is too complex to be listed. The main programming languages were Python and Java.The value delivered was outstanding. Data scientists were pleased by the fast deployments and development experience. Security and audit folks were overjoyed with the accountability provided. Business folks like the instant feedback and scores the models gave them.

Mobile application allowing the user to manage its wealth as well as webservices allowing wealth managers to give personalized care to their clients.Current status is: LEGACYTransformation and modernization of the technical stack.Helped consolidate the infrastructure by setting up on-premise Kubernetes. We industrialized the Python and Javascript deployments using CI/CD. Introduced coding norms and validation for Python projects. Supervised security audits.The stack was mostly Python and Javascript (ES6).

Designing and implementing a Geographic Information System (GIS).Current status is IN PRODUCTION.This solution helps with the integration and exploitation of the measurements done on the street relative to Fiber Optic cables deployments.The technical stack included Python (with various libraries), QGIS and PostgreSQL.Some change management was done to help convince and raise adoption. Operations cost and time was successfully reduced considerably.

Migration of the existing client facing website of France, Belgium and Luxembourg.Current status is IN PRODUCTION.Major migration of Adobe CQ5 to Adobe AEM. Stack wise, this is mostly a Java/JavaScript project.

Implementation of IFRS15 (accounting) related routines and set-top box events integration.Current status is UNCERTAIN.ETL operations through batch scripting.Stack wise, this is mostly Python running on a Cloudera cluster.

Migration of the existing Human Resources Information System (HRIS).Current status is probably LEGACY.The solution being migrated was HRAccess from its version 5 to 7.The stack included Cobol, Java and Python.The system was part of an offering given by the clients to its members.

Migration of the existing Human Resources Information System (HRIS).Current status is probably PRODUCTION.The solution being migrated was HRAccess from version 3 to 9.The stack included Cobol, Java and Python.This was a defense contract. This was a complete integration more than a migration given the number of major versions between the two systems.

Database migration factory developed internally for Accenture projects.Current status is: ONGOINGWrotes multiples modules to the existing monolithic solution. The first one (in Perl) was to help handle custom funtions written for domain logic when migrating data. The second one (in Delphi/Pascal) was a module to handle the conversion of data coming from mainframes (in EBCDIC) to MS SQL Server (in CP1252).The stack was mainly Delphi/Pascal Object and Perl.The migration was proposed to bigger projects to tackle the specific case of database migration. Often to clients of the Insurance and Banking industry (such as Allianz, Axa, etc…).