Hello, I’m Rémi Genet
I’m a Quantitative Researcher, PhD Candidate in Finance (expected completion July 2025), and Computer Science Lecturer at Université Paris Dauphine-PSL. My work bridges advanced quantitative methods with real-world finance applications, with a special focus on deep learning innovations for time series forecasting and execution strategies in cryptocurrency markets.
Research Interests
- Deep-Learning Integration in Finance
Rethinking execution problems by unifying prediction and optimization in a differentiable framework. - Time Series Modeling
Designing novel neural architectures (e.g., TKAN, TKAT, TLN) to capture time-series dynamics. - Open-Source Tools & Reproducible Research
Developing user-friendly packages and sharing research code to empower the community.
Current Positions
- Quantitative Researcher at Aplo (formerly SheeldMarket)
Conducting research on trading, execution, and risk management. I have developed a Python-based service that leverages deep learning models for real-time computation, which feeds into C++ execution systems. I’ve also built additional real-time C++ services for derivatives risk monitoring and liquidation algorithms. - Lecturer in Computer Science at Université Paris Dauphine-PSL
Teaching and developing course materials for:- Introduction to Python for Finance (since 2021)
- Object-Oriented Programming (OOP) in Python (since 2023)
- API Development in Python (since 2023)
- Deep-Learning for Finance (since 2025)
- Introduction to Python for Finance (since 2021)
Professional Experience
- Investment Manager (Apprentice) at ERAMET (Aug 2018 – Sep 2020)
Managed a multi-currency portfolio exceeding one billion euros, developed automation tools (Python, VBA, SQL) for asset management and reporting, and performed macroeconomic analysis.
Education
- PhD in Finance (Feb 2023 – expected Jul 2025)
Université Paris Dauphine-PSL
My research rethinks VWAP execution strategies on cryptocurrency markets through deep learning. Key contributions include:- Developing novel neural network architectures for time series forecasting (e.g., Temporal Kolmogorov-Arnold Networks and Transformers, and Temporal Linear Networks).
- Integrating path signature methods into neural networks via custom libraries (such as Keras Sig) and hybrid architectures (SigGate, SigKAN).
- Establishing a unified differentiable framework that extends from static to dynamic, multi-asset VWAP models.
- Advanced Master in Big Data & Machine Learning (2020 – 2021)
Télécom Paris - Master in Economics and Financial Engineering – Specialization in Quantitative Finance (272) (2018 – 2020)
Université Paris-Dauphine PSL - Bachelor’s Degree in Applied Economics & Financial Engineering (2017 – 2018)
Université Paris-Dauphine PSL
Technical Skills
- Programming Languages:
Python (expert), SQL (advanced), Bash (advanced), C++ (intermediate), VBA (intermediate), Java (basics), Rust (basics), R (basics), Matlab (basics) - Data & DevOps Tools:
Docker (advanced), GitHub Actions (advanced), poetry & pyenv (advanced), maturin (basics) - Languages:
French (native), English (fluent)
Personal Statement
I’m driven by the challenge of blending cutting-edge research with practical implementation. Among the languages I work with, Python stands out for its elegant design and logical APIs, which enable the development of robust, efficient solutions.
I conduct most of my research using Keras 3, valuing its backend-agnostic capabilities. I particularly favor the JAX backend for its innovative philosophy and performance benefits. My contributions—from novel neural architectures for time series to real-time trading services—reflect a broad commitment to open-source and reproducible research. I regularly share my code and develop user-friendly packages so that my work benefits the wider community.
I’m always open to collaboration or new opportunities. Feel free to connect with me on GitHub, LinkedIn, or check out my latest work on Google Scholar.