PROFILE SUMMARY

Passionate ML & Web engineer, skilled in Python, and its diverse set of ML frameworks (TF, PyTorch, JAX). Also skilled in JS/TS and web app engineering. Proficient in working with on-site and remote diverse teams.

Notable experiences:

• Being a core original contributor to the open-source library, tf-encrypted, for privacy-preserving machine learning in TensorFlow, which was accepted at the NeurIPS 2018 PPML workshop
• My research internship on Gounded language learning and modality fusion for which you can find the annotated slides here
• Founded a company in 2012 called Explee, an online platform for creating animated videos.

SKILLS

• Proficient in Python with experience in TensorFlow, PyTorch, and Jax (Local and Distributed Settings).
• Strong knowledge in infrastructure engineering (AWS, Docker, Linux) and management (CI/CD, git, etc.).
• Average knowledge of CUDA and Triton.
• Notion of C++ and Golang.
• Proficient in TypeScript/JavaScript and PHP.
• Fluent in French and English, with limited working proficiency in Italian.

RESEARCH & ENGINEERING INTERESTS

• ML scalability: Reducing the need for compute by improving data and models efficiency, and building scalable infrastructure and algorithms for ML research and deployments.
• Traditional-ML enhanced LLMs: especially LLMs enhanced with external planner, critics and verifiers.
• Explainable/Interpretable Machine Learning: developing tools to open-up the blackbox.
• Reinforcement Learning (RL), especially Multi-agent RL and LLM-enhanced agents to improve human-agent interactions.

WORK EXPERIENCE (ordered by date)

Led a remote team of 5 technical professionals at WoW Studio, a tech and media company that develops innovative experiences and products using blockchain technologies. (Ethereum L1/L2).

• Remotely led the development of the highly successful WoWG NFT collection.
• Actively participated in the developer hiring process by conducting interviews and creating technical assessments to evaluate candidates, ensuring a highly skilled team.
• Delivered multiple web application by coordinating and supervising developers' work.
• Established effective processes for task organization, promoting collaboration and transparent communication within the remote team. (main tools: Asana and Github)
• Served as the main contact for technical matters, offering guidance, support and code reviews to team members to enhance their skills and overall efficiency.

Built Leniax, a JAX library to apply state-of-the-art quality-diversity algorithms to the Continuous Cellular Automata Lenia:

• Explored the under-researched direction of QD algorithms in the Lenia domain, focusing on both algorithm design and implementation.
• Organized interdisciplinary research meetings with mathematicians and ML researchers.
• Published Leniax, a fully-fledged software to facilitate the use of QD algorithms on Lenia.

Research internship on Grounded Language Learning in simulated environments at ANITI (Artificial and Natural Intelligence Toulouse Institute):

• Conducted research under the supervision of Rufin VanRullen.
• Optimised a multi-agent simulation game to boost model training efficiency.
• Developed and trained several multimodal models for modality fusing (computer vision and natural language processing).
• Final presentation and annotated slides
• Project done using PyTorch: https://github.com/ruflab/soc

Core original contributor to the Open Source library for privacy-preserving Machine Learning in TensorFlow: tf-encrypted accepted at the NeurIPS 2018 PPML workshop. Associated research paper

• Contributed to the prototype that enabled Cape Privacy to secure its first seed round of funding.
• Developed discretized functions for neural networks, enabling training and inference to be conducted in cryptographic settings for enhanced privacy.
• Played a significant role in the Golang implementation of specific components within the project.

Explored various machine learning techniques, including:

• GANs for video compression (2019)
• Deep reinforcement learning with world models (2017)
• Unsupervised NLP (embedding, token prediction) and CV style transfer (2016)
• Blogger at https://medium.com/@morgangiraud (Main subject: ML engineering)

Worked as a freelance PHP/JavaScript developer, optimizing technical stacks, developping applications and structuring technical processes.

Taught two semester-long classes: PHP/MySQL and Algorithm/JavaScript, introducing students to programming languages, ecosystems, and basic algorithm concepts.

Co-founded the startup Explee as its CTO: an online platform for creating animated videos.

• Developed the entire technical stack for Explee.
• Built and managed a complex distributed web application infrastructure using AWS, Docker.
• Designed and developed the video rendering engine in JavaScript.
• Created computer vision algorithms for automatic drawing.
• Designed an managed the MySQL database.

EDUCATION:

Leading graduate school of engineering and part of the Institut Mines-Télécom, a top institute for Information and Communication Technology in France.

Intensive courses in advanced mathematics and physics, preparing for the highly selective entrance exams to the French "Grandes Ecoles."