About Me
Prior to the MFE, Loic studied statistics and economics at ENSAE Paris where he developed solid skills in stochastic calculus, probability, econometrics, and machine learning using different languages (Python, R and C++). He also holds an associate degree in law from Paris I Panthéon-Sorbonne University. His internship experience includes asset management, equity trading and quantitative research. While at the Dutch investment bank Kempen, he worked on several data science and machine learning projects to help traders identify signals. Then, in a quantitative research role at EY Advisory, he was responsible for developing and calibrating a local volatility model that improved option pricing precision as well as working on call-spread options P&L calculations and client engagements for major French banks. Loic enjoys exploring innovative methodologies with machine learning such as evolution strategies for trading and deep learning for options pricing. He also has a keen interest in systematic trading and quantitative strategies. Loic spends his free time reading economic news and studying geopolitics; he is also a longtime sailor and sailing instructor.
I am most skilled in: Python and Machine Learning
Projects
Evaluation of different models for price prediction
github.com/loicdiridollou/airbnb-nyc-price-predictionComparing different methods to predict AirBnB prices in New York City, regression, random forest, XGBoost, neural networks
Using a dataset from AirBnb Open Data, I conducted performance comparisons to predict the prices of stay. The best performance was achieved with a five Dense-layer network with regularisation and dropout with an accuracy of 70% while the XGBoost only achieved 68%.
Clustering customer data to identify and profile a certain type of customer that will conduct a precise banking operation
Using Santander customer data, the goal of this project was to explore different classification methods to predict a certain type of banking transactions. The best model was a combination of using XGBoost model and StratifiedKFold cross validation.
Transfer learning applied to bird recognition
The idea of this project was to explore the difference between the model that I would build on my own and a model that has already been trained. In the field of computer vision, the model needs a lot of data and training to perform well so without surprise the pretrained model (ImageNet18) surpassed the locally built model.
Computational analysis of different option pricing methodologies based on Monte Carlo simulation (academic project)
This project explored different variations of Monte Carlo simulations. The different analysis revealed that a Quasi Random Monte Carlo Simulation was the one that performed the best among the different ones (antithetic variables, Quasi MC, Random MC, …). This project was conducted with the Black-Scholes European Call closed form as a benchmark and two metrics were tracked for performance assessment: speed of convergence and variance of results.
First overview of computer vision and facial recognition mixed with my favorite TV series Suits
I started this project to introduce myself to computer vision. I am also a fan of Suits and I thought joining the two ends could be a great way to move forward with this project. I thus designed a face recognitiion using YOLO algorithm and open- cv. The idea was to identify the face in the image (treating the TV series as a sequence of images) and later identifying the faces of the character. It achieved a really good performance in recognizing faces with a 97% accuracy.
Experience
Helping the Risk Service Line serves its clients through the research and implementation of asset pricing models
I lead a project on researching and developing a model for local volatility in the field of Vanilla options. The project focused on taking in consideration the smile of volatility observed in the market, the evaluation of the model was conducted through the comparison of different pricing with generated and real data (to offer a wider understanding of the performance). Furthermore I also contributed to different client engagements to support the team and provide help on specific asset valuation of equility-linked swaps for a top tier French bank using multiple data sources.
Helping the Risk Service Line serves its clients through the research and implementation of asset pricing models
I lead a project on researching and developing a model for local volatility in the field of Vanilla options. The project focused on taking in consideration the smile of volatility observed in the market, the evaluation of the model was conducted through the comparison of different pricing with generated and real data (to offer a wider understanding of the performance). Furthermore I also contributed to different client engagements to support the team and provide help on specific asset valuation of equility-linked swaps for a top tier French bank using multiple data sources.
Helping the Risk Service Line serves its clients through the research and implementation of asset pricing models
I lead a project on researching and developing a model for local volatility in the field of Vanilla options. The project focused on taking in consideration the smile of volatility observed in the market, the evaluation of the model was conducted through the comparison of different pricing with generated and real data (to offer a wider understanding of the performance). Furthermore I also contributed to different client engagements to support the team and provide help on specific asset valuation of equility-linked swaps for a top tier French bank using multiple data sources.
Studying machine learning models to extract market signals in the field of life science company trading
As an Equity Trading Intern, my work focused on bringing data-driven approaches to stock picking and helping traders identify opportunities in the trading of life science companies. I was in charge of developing APIs, building data pipelines and asessing the performance of different methodologies to extract information from the financial market data. I worked a lot in collaboration with the Quantitative analysts to monitor the implementation and deployment of these different data-driven approaches.
Providing data-driven solutions to protecting portfolios from unwanted market moves
As part of the Credit Investment Grade team, I was in charge of creating a hedging tool based on market data to help portfolio managers manage the risk in their portfolios. I was also exploring portfolio profiling to support the growth in demand of SRI (socially responsible investment) and ESG (environmental, social, governance) labelled portfolios.
Education
UC Berkeley
Master of Financial Engineering
2020 - 2021
Top 3 MFE program, focusing on statistics, machine learning and applications in the quantitative finance world
As an MFE student at UC Berkeley, my classes focus on statistics, quantitative finance and I have elected the machine learning specialization to sharpen my skills in machine learning and deep learning, mostly applied to finance projects.
ENSAE Paris
Master of Statistics and Economics
2017 - 2021
One of the most prestigious Grandes Écoles in engineering in France, specialized in statistics and machine learning
I have chosen to specialize myself in data science and machine learning for finance. My classes focused on statistics, time series, probability and quantitative finance. I have conducted different projects in statistics and quantitative finance. I have worked with high-frequency data of the electricity spot market in Germany to study liquidity questions and predictability of price moves according to limit orders in the market.
Lycée Louis-le-Grand
Advanced Classes of Mathematics and Physics
2015 - 2017
Top 3 Classes Préparatoires in France, preparing for competitive examinations after two years of intensive maths, physics and computer science classes
Very intense syllabus in theoretical maths and physics, the syllabus also include classes in computer science and engineering science. My final project that was part of the competitive examinitions focused on modelling highway traffic to determine the right dimension of a toll gate to yield the best density of cars passing through the toll gate.
A Little More About Me
Alongside my interests in data science and machine learning some of my other interests and hobbies are:
- Sailing
- Photography
- Geopolitics