Yannick Le Cacheux, Developer in Paris, France
Yannick is available for hire
Hire Yannick

Yannick Le Cacheux

Verified Expert  in Engineering

Deep Learning Scientist and Developer

Location
Paris, France
Toptal Member Since
August 18, 2020

Yannick is a developer who holds two master's degrees and a PhD in machine learning. He has successfully carried out many ambitious data science projects for Fortune Global 500 companies as well as research institutions. Yannick has also authored several articles in international scientific journals, teaches machine learning in graduate classes, and is a co-author of a deep learning book to be published by the end of the year.

Portfolio

CentraleSupélec
Python, PyTorch, Machine Learning, Deep Neural Networks
Saint-Gobain Group
Artificial Intelligence (AI), PyTorch, Python, SQL, Deep Learning...
L'Oreal
Artificial Intelligence (AI), Deep Learning, Computer Vision, PyTorch, Python...

Experience

Availability

Part-time

Preferred Environment

Linux, PyCharm, Jupyter Notebook

The most amazing...

...image classifier I have developed uses text descriptions of classes instead of images during the training of the model.

Work Experience

Machine Learning Lecturer

2019 - PRESENT
CentraleSupélec
  • Created and taught the machine learning class for the master's degree in artificial intelligence.
  • Co-created and taught the deep learning class for the master's degree in data sciences and business analytics.
  • Supervised student projects and graded exams in machine learning in several other graduate classes.
Technologies: Python, PyTorch, Machine Learning, Deep Neural Networks

Lead Data Scientist

2021 - 2022
Saint-Gobain Group
  • Managed a team of 10 data scientists, providing technical guidance and expertise.
  • Overviewed and contributed to models for the optimization of the sales force's customer portfolio and store distribution.
  • Overviewed and contributed to models for sales forecast, prediction of supplier lead time, and optimization of store inventory.
  • Overviewed and contributed to models for client segmentation and profiling, business lifecycle detection, and estimation of clients' sales potential.
Technologies: Artificial Intelligence (AI), PyTorch, Python, SQL, Deep Learning, Machine Learning, Data Science, Deep Neural Networks

Senior Data Scientist

2020 - 2021
L'Oreal
  • Developed models to detect and evaluate clinical signs of aging (wrinkles, dark circles, etc.) from multispectral photos.
  • Created a fast and lightweight method to estimate the 3D shape of a face from single- and multi-view pictures based on statistical shape modeling.
  • Integrated the developed models on iOS with Core ML to deploy them in Lancôme points of sale.
Technologies: Artificial Intelligence (AI), Deep Learning, Computer Vision, PyTorch, Python, OpenCV, Generative Adversarial Networks (GANs), Machine Learning, Deep Neural Networks

Deep Learning Scientist

2017 - 2020
Commissariat à l'Energie Atomique (CEA)
  • Developed cutting-edge models to outperform all existing approaches in "multimodal" tasks involving both images and text in natural language.
  • Published several scientific articles in international peer-reviewed journals and conferences.
  • Contributed to the writing of a deep learning book as a co-author.
Technologies: Scikit-learn, Python, TensorFlow, PyTorch, Deep Learning, Computer Vision, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), GPT, Generative Adversarial Networks (GANs), Machine Learning, Deep Neural Networks

Data Scientist

2016 - 2017
AXA Group
  • Led the data science team in the deduplication and client knowledge project to identify links and patterns among millions of clients in multiple databases.
  • Designed predictive algorithms deployed in lead management information systems to optimize prospect conversion.
  • Developed a model analyzing and redirecting requests from incoming emails from clients.
  • Designed data science tests and conducted technical interviews.
Technologies: SQL, Apache Impala, Apache Hive, Scikit-learn, Hadoop, Spark ML, Apache Spark, Python, Machine Learning

Data Science Consultant

2015 - 2015
CGI Business Consulting
  • Developed big data and analytics products for large-scale unsupervised data visualization and clustering.
  • Supervised the deployment of large-scale flight-tracking software related to the passenger name record European directive.
  • Benchmarked many existing NoSQL databases and cloud platforms.
Technologies: Elasticsearch, Azure, Amazon Web Services (AWS), Tableau, Spark ML, Apache Spark, Cassandra, MongoDB, Scala, Java, Web Development, Machine Learning

Analyst

2014 - 2014
Goldman Sachs
  • Developed a tool to monitor activities on the futures trading platform as close to real-time as possible.
  • Analyzed past anomalies to predict most likely future malfunctions on trading platforms.
  • Tracked and fixed bugs on trading platforms using Jira.
Technologies: Groovy, SQL, Java

Generative AI for Creative Professionals

An app that I developed for Awen.ai. I was the lead data scientist on this project. The objective was to help creative professionals find ideas and iterate quickly on existing ones using generative AI. My work included modifying open-source image generation models, such as Stable Diffusion, and adding the ability to quickly integrate new concepts (objects, styles, etc.) within the model's knowledge base. I also integrated and fine-tuned large language models, such as Bloom and GPT-3.

Customers' Portfolio Optimization

A model that I developed for Saint Gobain. I was the lead data scientist for this project. The objective was to optimize the portfolio of sales representatives based on geographic criteria while balancing constraints (e.g., minimum and maximum size, revenue, etc.) per portfolio. While the problem formulated is NP-complete, a satisfying solution can be found using a greedy heuristic based on linear programming.

3D Face Estimation

A machine learning model that I developed for L'Oréal. I was the lead scientist on this project. The model's objective was to rapidly estimate the 3D shape of a subject's face from one or several photos. It consists of an iterative approach based on statistical shape modeling. First, key points are detected in the photo(s), followed by the orientation, scale, and more. Then, the most likely principal components are iteratively estimated until convergence, which is typically reached in four steps. The model was integrated on iOS with Core ML.

Zero-shot Image Classifier

https://tinyurl.com/ICCV2019-lecacheux
A deep learning model that I developed for CEA. A zero-shot classifier aims to recognize images of classes for which no training instance is provided, based only on semantic information. For example, even without ever seeing a tiger, most humans should be able to recognize one if provided with the information that "a tiger is a big cat with orange fur and black stripes on its back."

Zero-shot classifiers can be useful when one does not have training data for all classes.

I was the lead scientist on this research project aiming to improve existing zero-shot classifiers. I, together with two other deep learning researchers, showed that most existing approaches lacked desirable theoretical properties. More specifically, the usual loss functions do not enable the model to capture certain intra-class and inter-class structures. We provided novel theoretical results and developed a new model capable of outperforming all previous existing models.

Our proposed approach was published and presented at the 2019 International Conference on Computer Vision in Seoul, Korea.

Deduplication and Client Knowledge

A predictive model that I developed for AXA. The main objective was to identify clients with several unlinked entries in internal databases, for instance, because they independently purchased several insurance products. This is essential from a client's knowledge point of view.

At first glance, this is neither very difficult nor exciting: if two people have the same name and address, they are the same person. Problem solved. Except homonyms exist, addresses change, typos are made, and databases can be messy. And with millions of clients, a brute-force pairwise comparison is not an option.

Hence the need for clever predictive models to efficiently and accurately identify duplicates. I was in charge of the data science team designing the models and assessing associated risks.

The regularly updated information provided by these models is now a central component in many internal information systems.

The project was presented at the 2016 Viva Technology show in Paris.

Languages

Python, Java, SQL, Scala, Groovy, R

Frameworks

Apache Spark, Hadoop, Spark

Libraries/APIs

PyTorch, Spark ML, TensorFlow, Scikit-learn, Pandas, OpenCV

Paradigms

Data Science

Other

Deep Learning, Computer Vision, Natural Language Processing (NLP), Machine Learning, Artificial Intelligence (AI), Deep Neural Networks, GPT, Generative Pre-trained Transformers (GPT), Generative Adversarial Networks (GANs), Operations Research, Applied Mathematics, Web Development

Tools

PyCharm, Apache Impala, Tableau

Platforms

Jupyter Notebook, Linux, Amazon Web Services (AWS), Azure, Google Cloud Platform (GCP)

Storage

Google Cloud, Apache Hive, MongoDB, Cassandra, Elasticsearch

2017 - 2020

PhD in Deep Learning

Hautes Etudes Sorbonne Arts et Métiers - Paris, France

2014 - 2015

Master's Degree in Machine Learning

Georgia Institute of Technology - Atlanta, GA, United States

2012 - 2015

Master's Degree in Computer Science

Institut Mines Telecom Atlantique - Nantes, France

SEPTEMBER 2016 - PRESENT

Certificate in Data Science and Engineering with Apache Spark

edX

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring