Experience

Massive Entertainment - A Ubisoft Studio

• Sweden
• Computer Games
• 500 - 600 Employee

• Technical Lead

  • Dec 2022 - Present

  Ubisoft Scalar Project. Distributed High performance Microservices, C++, Concurrency, Persistence, Entity Synchronization.Technical Project Management

• Senior Online Programmer

  • Apr 2022 - Nov 2022

  Ubisoft Scalar Project. Distributed High performance Microservices, C++, Concurrency, Persistence, Entity Synchronization



Amadeus

• Spain
• IT Services and IT Consulting
• 700 & Above Employee

• Senior Software Development Engineer - Senior Expert

  • 2021 - Feb 2023

  Micro Services (Reactive, Java, Quarkus, JAX RS, CDI, Openshift, Kubernetes) (as Senior Expert)

• Senior Software Development Engineer - Senior Expert

  • 2013 - 2021

  Cloud Journey RBox (GCE, AWS, C++, Python, Go) (as Senior Expert): In this team, we started innovating for Amadeus. We were the first to migrate completly a such high transactions per second application in the Cloud (GCE and AWS). I did project coordination, initiating internal open source. I implemented the first Elasticity Application in Amadeus (30% bill reduction: hundreds of K euros per month). In this Cloud Journey, we put in place the DevOps culture. What we develop we also operate.Cloud Journey Perfect Stagging (GCE) (as Senior Expert): This new version of our stagging enviroment now includes Database performance QA and Functional QA. It was the first application in Amadeus with such level of qualification. Separating read and write flows in your appplication enabled read flows to benefit from such QA level. The Write flows are still a difficult problem for QA.Cloud Journey Architecture for Amadeus Global Migration to the Cloud : (AWS, Azure, Python, Go, Openshift v4) (as Senior Expert). Our full application portofio migration to the cloud started with our team of Exploration and Design. We designed around Infrastructure as code, Network, Load Balancing, DNS, Openshift Cluster v4. First on AWS and then on Azure (When Amadeus signed a partnership with Microsoft). Working with MSFT Architects we built the Cloud Platform Architecture for Amadeus (Hundreds of applications to migrate to the Cloud). This is a major project in the Company. In addition to this System and Cloud Architecture, I also put in place a new Software Development stack for infrastructure, with Python Pulumi first and then Terraform. The development cycle was complete with Source Control, QA, CICD. Show less

• Software Development Team Lead

  • 2009 - 2013

  APE Design : (C++, Design, Team Leading)Again, I did a full Design from scratch, inspired from the Altea Design, but with the lessons learned. Inputs/outputs de-coupling re-visited. It was the first big application migration from TPF assembler mainframe to Linux C++. First High TPS application (Low Fare products with around several thousands transactions per second). I did asynchornous processes and simple generic state machine design. But, with evolution, I saw how over-engineering can hit you: The evolution brang a lot of complexity in this "simple" and generic state machine. The generic engine became too complex to maintain compared to several different processes using common modules.I built a strong team spirit and engagement.Optimization Task Force on APE (C++) (as Technical Expert)Analysing the code with profilers like Callgrind or perf, I led a task force and implemented optimisations bringing between 30% to 5% CPU gains (first big wins quickly reducing to only small wins in the pipe). I also co-worked on low level optimizations like Factory pooling with objects locality (between 5%-10% optimization).APE Staging : (C++, QA) (as Technical Expert):I fought and brang some Innovation in the QA field of the APE application. Our performance testing enviroments were irrelevant. We got a lot a false positive. So, I pushed forward this idea of stagging environment for performance and memory QA. Perf tests was done on PRD isolated nodes using shadow traffic (duplicated traffic from PRD, ignoring replies). With such environment, the performance and memory qualification kept relevant, without impact on customers. But still, the Database was NOT qualified with the right throughput of queries as the Database on Production was still centralized.TIA refactoring (C++, QA) (as Technical Expert)Refactoring of an application Foundations of the code. Extensive QA allowed me to bring this major change with 0 issues on PRD. Kind of achievement for a developer. Show less

• Software Development Engineer

  • 2001 - 2009

  Altea Backend Design and Implementation : (C++ servers, Java for UI and Oracle). We started from scratch. I Designed a full layered software: with inputs/outputs decoupling, Memory Management (no smart pointers), Fields (Low level DB compatible attributes with or without polyphormism), Database access control. I writed documentation, samples, trainings, I presented the design, the phylosophy, in order for hundreds of developers, to start implementing this major product in the company. I implementated the low level layers like Memory Management Fields, Database access.I got experience and evolution of such Design over 10-15years.Oracle Optimisations : (Oracle).I learnt to often hint the Oracle Optimizer in order to ensure control and stability of the execution plans. I had to control such plans, especially the sequence of tables in joins, and the indexes used for such sequence. This allowed to bring robustness and stability on our Production. It also allowed to get 30% to 50% performance gains. But, there's always a great debate between using or not the Oracle Optimizer (to hint or not to hint, that's the question). Show less



ACT - Mysis

• Software Development Engineer

  • 1996 - 2001

  Portfolio Applications : (PowerBuilder, Sybase) I developed Financial Portfolio Management Applications. C++ Backends refactoring : (C++, Design) We refactored the current C++ servers. We got a lot of core dumps, and memory issues. Based on a Factory of objects we brang robustness and stability. Core dumps disapeared. Database optimisation : (Sybase) I ramped-up on the concepts of execution plans, table joins sequence, indexes. With such knowledge, I was able to optimise Sybase queries. The gain can be important. This also bring stability in queries performance. Which is an important aspect of the Production. When using relational Databases, the performance optimization of processes is often linked to the Database one. I created trainings to share such a knowledge inside the company. Show less