Architect and CoreTech developer of a media analytics solution (press, web & social media)Designed the Spotter architecture capable of handling millions of multilingual documents per day in near real time.I wrote Spotter’s query engine on top of low-level Lucene components. The query engine (aimed toward expert users) is able to do complex nested span queries (word/sentence/paragraph "slops"), query part of speech and syntactic relations (with lucene payloads).I developed a NLP rule engine (similar to Stanford TokensRegex) on top of Gil Francopoulo's TagParser (syntactic parser). This engine is used to extract quotes (direct and indirect) in both French and English.Designed a contextual influence score on Twitter content. Study and prototype with R and implementation was done with MapReduce on top of Mongodb.

Studies and audits on software engineering methodologies at La Poste in the scope of a CMMI certification. Contributed to Apache Continuum a continuous integration platform. I wrote studies and proposals to move from Synergy to free software configuration management solutions like Git.

I was part of a team of 10 developers to rewrite the BRLe (farm water seller) back-office written in DataCom/Cobol into J2EE technologies (evaluated at 2500 working days). I took tech lead responsibilities during the last phase of the project (acceptance testing and deployment). Beside this big development project I also provided IT consulting for various other customers.

I worked on the architecture to collect data from various press providers (AFP, Reuters, Comtex...). I developed back-office components with Perl and front-office with Apache Cocoon.

I worked on an automatic installer to deploy disk-less GNU/Linux desktop solution for small business. This job included various tasks, including Perl and C development, automatic GNU/Linux distribution configuration and packets management...

I developed a domain reseller website written in Perl.