Matthieu TOURNE
Software Engineer, CloudFlare, Inc.
I have various interests in CS from distributed architectures, large scaling, to security (especially in wireless), system programming, and even web programming.
I enjoy mental challenges, and learning new possibilities to make things more efficiently.
Skills :
- Programming Languages: C/C++, Python, Java, PHP, Javascript, OCaml.
- Networking : IP (v4/v6), TCP/IP, 802.11 (WiFi)
- Systems : Linux, Mac OS X, OpenBSD, Windows.
School projects :
- Presentation of the WiFi security (WEP/WPA) and demo implementation of “caffe latte” an innovative attack targeting the client instead of the AP.
( http://mattourne.blogspot.com/2008/10/caff-latt... )
- Deployment of a lab IPv6 only network to analyze Windows XP IPv6 support, set up of NAT-PT for translation on adjacent v4 networks, and AD using samba.
- Kaneton : Micro kernel for x86 architecture, implementation of the memory management, interrupts and scheduler. ( http://en.wikipedia.org/wiki/Kaneton )
- Unix Shell, full featured clone of “bash”. Portable across unices, written in C.
- Tiger programming language compiler, using lex/yacc and C++.
Languages:
- English : fluent
- French : mother tongue
64 contactsWorking on the core of CloudFlare CDN technology: Reverse Proxy, Caching, Web performance, html rewriting, WAF.
This includes extending and modifying for our needs the webserver and reverse proxy NGINX, an opensource project written in C by Igor Sysoev.
I also run benchmarks and write tests for our infrastructure.
2009 - 2009Work on QualysGuard web application.
Development in php, and javascript (jquery).
Participating in creating a new FDCC/SCAP module to comply with federal requirements.
2008 - 2008Site Reliability Engineering of the Ads Systems
Work in a team maintaining a very large and "sharded" MySQL Database.
- Development of a SQL parser, to detect harmful schema changes on the production (sql_lint). Design of a mechanism to write new detection rules.
- Contributed to modules of a MySQL "copilot". A program capable of "fixing" the DB in real time, by pushing configuration changes, killing long running queries based on patterns, or taking out the instance from the serving pool if the binlog replication falls to far behind.
- Learned with great interest about vertical distributed architectures on commodity hardware, such as Bigtable or HBase (Hadoop)
2004 - 2004Monitoring de Production Industrielle à l'aide de clusters haute disponibilité (IBM High Availability Cluster Multiprocessing for AIX 5L)
Mission :
Outil de monitoring des serveurs et bases de données.
Assistance à l'administration réseau.