About
Activity
2K followers
-
Thomas Leplus shared thisLooking for a new year's resolution, you can visit this site from your home Wi-Fi (and your office, your relatives and friends homes...): https://lnkd.in/en7eMeia If you don't see a green checkbox, read the explanation why carefully. Otherwise you're done until next year. Easy!
-
Thomas Leplus shared thisHi everyone, You or someone you know is looking for an AppSec role in Paris, look no further 😉 https://lnkd.in/eXMJbh5u Don't hesitate to ping me if you have any questions regarding the role. Happy holidays 🎄
-
Thomas Leplus shared thisAn unauthenticated remote code execution in front facing frameworks like React and Next.js should send shivers running down your spin. Please pay attention if you use these technologies. https://lnkd.in/eQEVPbg6React2Shell (CVE-2025-55182): Critical React Vulnerability | Wiz BlogReact2Shell (CVE-2025-55182): Critical React Vulnerability | Wiz Blog
-
Thomas Leplus shared thisThink you know a valid e-mail address when you see one? Think again. Here is a fun little quiz to test your RFC chops: https://e-mail.wtf I scored just 13/21, barely above a 50% guess 😳 And you? On a more serious note, we keep asking developers to validate user inputs to prevent injection attacks and we wonder why it's not working. Well here is why. Today's standards must make security practical otherwise we still won't stand a chance years from now when these standards will be widely adopted.
-
Thomas Leplus reposted thisThomas Leplus reposted thisThe first OWASP AppSec Days France is a wrap! 🎉 From meeting my podcast guests in person (François Proulx) to diving deep into great conversations, here's what stood out: - Software Supply Chain Security is hot 🔥 From Roni Carta’s keynote to François Proulx's insightful talk, and the many conversations in the halls about recent attacks, it's clear that supply chain security is top of mind for everyone. - It was inspiring to see so many developers/security champions attending the conference. We’re definitely headed in the right direction, but there's still more work to do to engage even more of them :) - Presentations that share direct, real-world experiences were incredibly valuable for the audience, sparking lots of questions! And congrats to Nohé Hinniger-Foray on his first talk about Escape's upcoming research report about the state of vibe-coded applications, vulnerabilities and secrets we found 👀 Stay tuned in the upcoming weeks! Big thanks to all the speakers, volunteers, sponsors, and everyone who made this event a success. A huge thank you to my partners in crime: Yvan PHELIZOT, Marine du Mesnil, Paul Molin, Didier Bernaudeau! Thomas Leplus, Roni Carta, Edouard Viot, François Proulx, Dwayne McDaniel, Nohé Hinniger-Foray, Nourredine Khadri, Jérôme Dupuis, Adeline VILLETTE, Daniel Garnier-Moiroux.
-
Thomas Leplus shared thisMany thanks to the OWASP France chapter team for organizing the first edition of the OWASP AppSec Days France conference [1] and for letting me volunteer to help. If you missed it this year, mark your calendar for next year; it's definitely worth it. And in the meantime, there are also monthly meetups [2] so if you are into AppSec and you live anywhere near Paris, you have (almost) no excuse 😉 [1] https://lnkd.in/enjag8vR [2] https://lnkd.in/enNtPe6R #owasp #owaspfrance #appsec
-
Thomas Leplus shared thisSecure code in a world of generative AI Generative AI already had profound impacts on software engineering. Today’s programmers use LLM-based coding assistants to “fill the blanks” in their code. AI chatbots produce code snippets helping developers solve their problems. More advanced tools can generate the scaffolding to architect new applications. And this is only the beginning. One of the few things that are certain about the future is that any prediction that you make will sound naive in retrospect. But it's easy to picture a future where human languages become the 6th generation programming languages used to create and maintain code, and today’s programming languages become mostly hidden “under the hood”. This is nothing new, modern programming languages started replacing assembly languages in the late 50s just as assembly languages displaced machine languages before them. Each generation provided more complex abstractions to express more advanced concepts, just as human languages did. So it seems quite natural that at some point the two should meet in their evolutionary paths. But coming back to the present day, code generated by LLMs is steadily reaching the level of a university student. The code snippets produced by LLMs often resemble the answers one would find on a site like Stack Overflow or GitHub. This is to be expected since this is exactly the kind of data sets that the models are trained with. Sadly code security is often the last concern (at best) in academia or on coding advice websites. So the same is reflected in the code produced out of the box by LLMs. If you explicitly ask a model to generate secure code, the results are more encouraging, especially with newer "reasoning" models. However, experience shows that it is a Sisyphean effort to ask developers to follow secure coding best practices. Why then should we expect them to request secure code from their tools? If it is easy, we might see some improvement, but if we want to eradicate insecure code, we must aim higher. Tools must produce secure code by default. Just as you don’t need to ask for safe brakes when you buy a car, why should you be required to ask models for secure code? It should be an implicit requirement enforced as a company policy, not by individuals. The ongoing paradigm change in how software is made is a once-in-a-generation opportunity to turn secure code into the norm. I hope that as an industry we do not miss this narrow window. If you are currently looking at which code-assisting technology to use, I strongly urge you to make security a part of the conversation and to involve your security team in the selection process as early as you can. If you are the security team, you should embrace this change as an opportunity and use your existing tools (security code reviews, SAST, etc.) to evaluate the tools and challenge vendors to do better. Whether you agree, disagree (or both), please feel free to share your thoughts in the comments. Cheers!
-
Thomas Leplus liked thisThomas Leplus liked thisHonored to be recognized for 20 years in FinTech. I appreciate the recognition Nasdaq and the many colleagues who have made this journey incredible. Looking forward to what’s ahead.
-
Thomas Leplus liked thisTrès honoré de figurer dans le livre des Success Stories des 90 ans de l’EFREI - Grande école du numérique, en tant que cofondateur et CTO de ORMEX , une startup engagée dans le développement durable. https://efrei90ans.fr Merci à l’EFREI - Grande école du numérique pour cette reconnaissance. 🚀 L’aventure continue, avec sens et impact. — Very honored to be featured in the Success Stories book celebrating the 90th anniversary of EFREI - Grande école du numérique, as Co-Founder and CTO of ORMEX , a startup committed to sustainability. https://efrei90ans.fr Thank you to EFREI - Grande école du numérique for this recognition. 🚀 The journey continues, with purpose and impact. #EFREI #Ormex #Alumni #SuccessStories #CTO #Sustainability #TechForGood https://lnkd.in/eYzncMUbThomas Leplus liked this🎉 L'EFREI - Grande école du numérique fête ses 90 ans ! Depuis 1936, nous formons des générations d'ingénieurs, d'innovateurs et de décideurs qui ont contribué à transformer le monde du numérique, la technologie et l'ingénierie. À l'occasion de cet #anniversaire, découvrez nos Success Stories : 🔹 Des #alumni emblématiques, acteurs majeurs de l'industrie et du #numérique 🔹 Des enseignants passionnés qui ont inspiré des générations par leur #pédagogie 🔹 Des avancées technologiques historiques, comme le Micral, 1ᵉᵈ micro-ordinateur au monde 🔹 Une communauté unie par les même valeurs : l'exigence, l'audace, l'ouverture et l'#agilité Merci à l’ensemble de notre communauté : #enseignants, doctorants, étudiants, #alumni, entreprises partenaires et collaborateurs pour ces 90 années d’innovation, de transmission de savoir et d’audace. C’est grâce à vous que l’EFREI - Grande école du numérique continue de grandir et d’inspirer ! 👉 Découvrez les portraits de femmes et d'hommes qui ont contribué à cette histoire riche dans notre livre Success Stories – spécial 90 ans de l'Efrei, disponible via ce lien : http://efrei90ans.fr/ Frédéric Meunier Pierre-Olivier Colliard Julien Renard 🔒 Philippe V. Arnaud MATHIEU Florent Cotton Laurent Bocquet Benjamin Roecker 💡RELAVIL Bellamy RODFORD Mick DUPONT Amaury PAQUIS-THONAT Frederic Naude Alexis Trancart Dieydi N'diaye Anna Martin Ketsia Nlandu Wa Kabila Edouard ARNAUD💡 Véronique Trançon Jean-Francois Lambert Charlotte Davain Caroline COIGNARD Bertrand GOBERT Hélène Torri Melissa K. Charlotte Bassot
-
Thomas Leplus liked thisThomas Leplus liked thisCOOs, l'automatisation sans gouvernance n'est pas une audace, c'est une vulnérabilité. 🚨Il est temps de poser les bases solides : politiques claires, gestion des cas, pistes d'audit et garde-fous indispensables.Chez Pega, nous savons que la vraie innovation repose sur le contrôle rigoureux des processus automatisés.Notre plateforme low-code et notre intelligence artificielle intégrée garantissent des transformations digitales sécurisées et efficaces.Ne laissez pas votre entreprise exposée, construisez la gouvernance dès aujourd'hui pour maîtriser le changement avec confiance. 💡La passion pour la technologie de pointe est notre moteur, et elle vous propulse vers l'excellence opérationnelle.Agissez avec audace, maîtrisez l'automatisation, et gagnez la confiance durable de vos clients. 🔥#TransformationDigitale #Automatisation #Gouvernance #Innovation #Leadership
-
Thomas Leplus liked thisThomas Leplus liked thisWishing you all peace and joy this holiday season and a happy 2026 from all at CFM.
Experience & Education
-
Capital Fund Management (CFM)
View Thomas’s full experience
See their title, tenure and more.
or
Volunteer Experience
-
Collection volunteer
The AIDS Trust of Australia
- 2 months
Health
I raised funds for HIV/AIDS organizations by collecting money in Sydney, Australia.
Publications
-
From the RealWorld to Real Words: the METEO case
EAMT-05, Budapest
Machine Translation (MT) is the focus of extensive scientific investigations driven by regular evaluation campaigns, but which are mostly oriented towards a somewhat artificial task: translating news articles into English. In this paper, we investigate how well current MT approaches deal with a real-world task. We have rationally reconstructed one of the only MT systems in daily production use: the METEO system. We show how a combination of a sentence-based memory approach, a phrase-based…
Machine Translation (MT) is the focus of extensive scientific investigations driven by regular evaluation campaigns, but which are mostly oriented towards a somewhat artificial task: translating news articles into English. In this paper, we investigate how well current MT approaches deal with a real-world task. We have rationally reconstructed one of the only MT systems in daily production use: the METEO system. We show how a combination of a sentence-based memory approach, a phrase-based statistical engine and a neural-network rescorer can give results comparable to those of the current system while offering a faster development cycle and arguably better customization possibilities.
Other authors -
-
The long-term forecast for weather bulletin translation
Machine Translation
Machine Translation (MT) is the focus of extensive scientific investigations driven by regular evaluation campaigns, but which are mostly oriented towards a somewhat particular task: translating news articles into English. In this paper, we investigate how well current MT approaches deal with a real-world task. We have rationally reconstructed one of the only MT systems in daily use which produces high-quality translation: the Météo system. We show how a combination of a sentence-based memory…
Machine Translation (MT) is the focus of extensive scientific investigations driven by regular evaluation campaigns, but which are mostly oriented towards a somewhat particular task: translating news articles into English. In this paper, we investigate how well current MT approaches deal with a real-world task. We have rationally reconstructed one of the only MT systems in daily use which produces high-quality translation: the Météo system. We show how a combination of a sentence-based memory approach, a phrase-based statistical engine and a neural-network rescorer can give results comparable to those of the current system. We also explore another possible prospect for MT technology: the translation of weather alerts, which are currently being translated manually by translators at the Canadian Translation Bureau.
Other authors -
-
Weather Report Translation using a Translation Memory.
Association for Machine Translation in the Americas
Weather Report Translation using a Translation Memory.
Thomas Leplus, Philippe Langlais and Guy Lapalme.
Association for Machine Translation in the Americas (AMTA-2004)
Georgetown University, Washington DC (USA), SeptemberOther authors -
Languages
-
French
Native or bilingual proficiency
-
English
Full professional proficiency
-
Spanish
Professional working proficiency
-
German
Elementary proficiency
Organizations
-
Apache Foundation
Contributor
-
OpenSSF (Linux Foundation)
Contributor
-
OWASP
Member
Recommendations received
-
LinkedIn User
7 people have recommended Thomas
Join now to viewOther similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content