Hello there! 👋
I am currently working as a research engineer at SUMMIT, at Sorbonne Université.
I am mainly contributing to the DIOPTRA research group, working on internet measurements and the Paris Traceroute tool more specifically.
We aim to ship significant improvements to the traceroute
tool and provide a more exhaustive picture of load-balanced routes.
I am actively engaged in research as well as development in Python and Rust.
See this article for a description of my contributions.
Other current projects include developing VR Experiences and data management apps for medical research at the ICM (Institut du Cerveau).
I am mainly using Unity, FastAPI and Vue.js for these projects.
My personal areas of interest include cryptology, natural language processing, teaching, music.
You can find my one-page CV here.
2023 — ongoing
2023 — ongoing
2023
2019 — 2023
2017 — 2019
2016 — 2017
2023 — ongoing
I am leading several applied research projects at SUMMIT for teams at Sorbonne Université and outside. SUMMIT is an engineering unit at Sorbonne Université that helps research labs through partnerships with the industry at large.
At LIP6, I am currently rewriting fast-mda-traceroute, a tool that produces exhaustive maps of all network paths between a source and a destination address.
This tool aims to be a replacement for traceroute
which does not account for multiple paths induced by load-balancers for instance.
My contributions include the rewrite of the combinatorics formula, extensive testing with real target destinations, putting together a known topology testing framework that simulates probing and general development.
Though most of the development so far is made in Python, work has already started porting the project to Rust.
2023 — ongoing
I am leading several applied research projects at SUMMIT for teams at Sorbonne Université and outside. SUMMIT is an engineering unit at Sorbonne Université that helps research labs through partnerships with the industry at large.
A project from the ICM aims at understanding the mechanisms behind the freezing of gait symptom of Parkinson’s disease patients. I am working on a VR video game that must trigger this freezing in a controlled environment. I am also working on corellating data from the game, motion capture sensors, and eye tracking in order to provided actionable data for researchers.
Development is done in C# / Unity for the VR part, Python / Matlab for the post-processing.
2023
After leaving the public sector I briefly joined a cyber-security start-up specializing in secrets discovery in code repositories. I learned useful techniques in Python Django development as well as a modern development and deployment framework.
I contributed in the evaluation of Infrastructure-as-Code (IaC) configuration, such as Terraform.
My other main contributions were in Prompt Engineering, by de-mystifying this new discipline and producing leads and proof-of-concepts for using LangChain and the OpenAI API in production.
2019 — 2023
I was appointed technical leader of my team in late 2019. I managed to double the team size, going from 4 to 8 people, recruiting a variety of profiles. I contributed to steer the team toward a more data-driven approach, while keeping a strong and open experimental team spirit.
During this period of growth, we streamlined the onboarding process, as well as the internship program. I supervised 10 student internships, and 5 employee internships. This was achieved through regular internal communications and multiple external meet-up events.
As with any long standing activity, we had to deal with large amounts of legacy code. Highlighting multiple concepts in a clearer fashion helped us clear away a lot of clutter. We effectively purged large amounts of code that weighted the team down.
The largest code port I supervised consisted in tens of thousands of lines of Perl code, that we successfully ported to Python. This was achieved by empowering the right developers, and balancing this relatively tedious task with more expressive projects.
Covid was an opportunity to strengthen self-training. Currently, we are one of the few teams that regularly work remotely. Responsibility and accountability lets us be very autonomous in that regard. Ambitious training and research let us participate in multiple research seminars, that I coordinate.
We collectively trained on reinforcement learning, through a MOOC.
During that time, I also had to manage several difficult situations, Covid being one of them. One inter-personal conflict also arose, and together with my team, we managed to reach a satisfying resolution.
With clear annual objectives, the team continued to deliver robust operational results, year after year.
Leveraging our strong autonomy, I enabled every agent to express themselves through their work, in line with our overarching objectives.
2017 — 2019
After working at Thales, I worked at the French Ministère des Armées (Ministry of Defense) on several applied research topics. In team of three to four people, I made significant theoretical and practical contributions that directly impacted operational successes.
Along with the team, we strived to industrialize more and more of our codebase (C, Python), and bring up a micro-service based architecture.
This transition was a fascinating time. When I arrived, we were using Subversion and VMs; two years later, GitLab, CI/CD, docker, Kubernetes, code quality metrics, monitoring, efficient logging, …
Novel ideas we pushed forward at the time were NLP and neural networks applied to our field, as well as new finite automata implementations. As a team, we trained on those topics and gained significant expertise.
2016 — 2017
I developed software to secure satellite communications, with a focus on rigorous testing. For two years, I acted as the main reference point on cryptography for my team.
During development, I designed novel automated testing methodologies (Python) while navigating and patching legacy code (C, C++).
Workflow was well established and I was lucky enough to learn Kanban methodologies in an industrial context. The two projects I took part in were a success and are still used in production today.
In a primarily research-oriented activity, long term vision is primordial. I strive to describe clear objectives, relying on the team self management when possible. I regularly assume the role of the client, pushing for quality, even when facing short term milestones.
With dozens of hiring interviews for a wide array of technical profiles, I have a firm grasp on how to evaluate candidates efficiently: I doubled my team size, from 4 to 8 members.
I conducted performance reviews regularly with my team, paying attention to personal wishes and initiatives.
I started multiple transformative projects, regarding development or infrastructure. We successfully conducted large code refactors without impacting operational imperatives.
See my currrent position for more details.
With a robust theoretical background, I co-designed and implemented secure key management for satellites. Recently, I have been interested in quantum computing and post-quantum cryptography.
Facing emerging challenges, I trained on specific topics and managed to assemble a team of experts. I recently completed the reinforcement learning specialization from the university of Alberta, and encouraged non-specialists in the team to train as well.
For massively parallel computations, I contributed to new meta-heuristics and continuous model deployment mechanisms. I designed, developed and proved several novel algorithms on finite automata and graph pathing.
Among other projects, I designed and developed the main library used to this day in my team. I always strive for great documentation and extensive testing, leading to an easy onboarding for our new team members.
With some experience optimizing C programs, I am now able to understand the challenges of high performance computing, and work with engineers designing and finding solutions.
I recently fell in love with Rust and spent a significant amount of time completing programming challenges in that language.
Pushing modern technologies in my current team, I acquired a firm grasp of the benefits of a microservice environment. I supervised and contributed to the transition from monolithic applications to a more manageable service based architecture, with some serverless parts.
As an early adopter of git in all my previous positions, I helped many of my colleagues with their first steps. I experimented with several forms of code review, finding which fits the best for individual characters.
Initially a vim aficionado, I came to enjoy the benefits of full-featured IDEs.
Internship at Thales Communications and Security on Securing biometrics in a chip-less ID.
Masters degree in Mathematics, Computer Science, and Cryptography, at Université de Paris
Research project: Reed-Müller error correcting codes
Bachelor’s degree in Mathematics and Computer Science, at Université de Paris
I would love to hear about new opportunities, regarding technical management, or development in an exciting scientific environment.
I’m living in Paris, France.
My current local time is .