
Graduating from engineering college in 2000, JB joined Margo in 2010 after spending the first part of his career in industry and at software publishers. This week, he starts his third mission for Margo, as the Team Leader of a team in charge of implementing the ActivePivot solution for one of our clients. He is equally well-known in the office for regularly hosting the Coding Dojo session.
Can you briefly outline your career before joining Margo?
I started at a consultancy where I honed my skills in OO languages, I worked in different sectors and in companies of various sizes, from a medical start-up to EDF. After 4 years, I went to a publisher, where I stayed for 6 years. I principally worked on problems around architecture, databases, GUIs etc.. That was when Scrum arrived and I was able to take part in setting up Agility for them. Up to then, I had worked in C++ and Java. I got to know Margo through a friend who co-opted me. I particularly liked the focus on quality, not only around clients but also consultants: their management and their missions.
What about your first two Margo missions?
I started with just over 3 years at Société Générale CIB on a very technical Java project. I had to work on a bunch of servers that aggregate and redistribute data, with constraints on bandwidth and response times. I also took part in setting up the Scrum method there. I had the job of Scrum Master and Team Leader.
Following that I worked for around a year and a half at BNP on a program which creates and books derivatives. There were a lot of interesting topics. One example was the creation of a non-regression system to prevent program outputs from being changed between two versions and to put in place a fail-fast system.
Today you’re starting a new mission with us. What sort of subjects have been suggested to you?
I had interviews for five different subjects at FIBs and at an Asset Manager. Most are fee-based but one is fixed-price. I had positive feedback from four of the clients. I chose a mission based around ActivePivot. It was hard to choose, but I had put ActivePivot at the top of the list. Not just for the interesting practical environment – the risks – but also for the technical constraints which stem from it: a huge amount of data… I’m talking 7 terabytes in memory! With OLAP solutions, loading data and running queries is usually slow. With ActivePivot you can load everything to memory and get very short response times, less than a second, even for complex queries. That interested me and it’s cutting-edge technology. Plus, I had a good rapport with the project manager and the team was extremely flexible, which never does any harm.
What are the specific challenges in the coming few months and years?
ActivePivot is a solution where you work in N dimensions. It’s quite distinctive. There aren’t really any other solutions like it on the market. And you have to train in the technology. As part of that, I did a month’s full-time training on the Quartet FS site before starting my mission as Team Leader for the team implementing this solution.
What’s in this solution for the FIBs, do you think?
I’ve known other FIBs where the data routed via flat files which were re-worked with scripts for feeding into a database. In practice, that meant that users suddenly had access to the previous day’s data, and there was a huge time delay. It wasn’t very user-friendly either – if you wanted to get different information or to process the information in another way, you had to re-work a specific function and be shared for 24 hours. With ActivePivot there’s no such worry… on the other hand, it is true that implementing and using the solution is complicated and you need training!
How would you describe the possible development of this solution?
It’s not used everywhere yet, especially because Quartet FS is working with proprietary rather than open source code, and it’s still a niche market. But it’s looking good for the future.