Freelancer profile translated to English.

Description

I am a Data Scientist using Python and Spark ML, co-founder of Effets Mesurés (Online custom clothing sales at mue-store.fr), for which I developed the measurement-taking algorithm.

I am accustomed to producing quality code thanks to the requirements and rigor of my engineering training at EPITA (SCIA 2018 class, Data Science/Data Engineer)

I can work on topics in Python and Keras/Tensorflow, Pytorch, Vue.js; mainly on Computer Vision but also NLP topics.

I can handle different stages of the artificial intelligence pipeline:

Needs Analysis:

- Dialogue with the client to clearly define and express the need

- Translation into a machine learning pipeline meeting the selected criteria

Data Retrieval:

- Scraping (selenium, scrapy)

- Various API calls: Google Sheet, Pipedrive, PubMed, etc.

Data Processing:

- File reading

- Data cleaning and processing (Pandas, Numpy, etc.)

- Data storage and transformation using SQL or NoSQL databases

Data Visualization

- Python visualization libraries: Matplotlib, Seaborn, Plotly

- Dashboard generation with PowerBI (Microsoft) and QuickSight (AWS)

Data Analysis, Prediction, and Classification

- Preliminary data analysis

- Training of machine learning or Deep Learning models

- Use of ChatGPT for Question Answering with a vector database (FAISS, Pinecone, etc.) or to link it to an internal API (product recommendation, etc.)

- Choice of performance metrics and performance analysis to validate the model

Creation of containerized applications (Docker)

- API creation with Flask, FastAPI

- Application creation with Streamlit and Flask, deployment on AWS or GCP

Industry field of expertise

Languages

French
Native or bilingual
English
Fluent
Spanish
Conversational

Workplace preferences

Can work on-site

Paris (up to 10km)

Cleed
Lead Developer & Data Scientist
E-COMMERCE
October 2022 - July 2023 (9 months)
My mission at Cleed consisted of three main achievements:

- Development as lead developer of the backend for client Cabinet Septembre for an administrative document management platform. I was responsible for the design, implementation, monitoring, and team management for the backend. All of this followed an agile methodology.

- Design, architecture, lead, and development for a second version of the chatbot solution offered by Cleed.

- Integration and improvement of NLP components for the chatbot, particularly through OpenAI APIs. Linking the LLM model with an internal API to connect the model to a database.
Deep Learning Chatbot NLP Data Science Pytorch Python Lead Developer Back-End Development Typescript AWS Web Architecture Agile Methodology
Zenly
Style transfer from Zenly design
INTERNET OF THINGS (IOT)
April 2022 - July 2022 (3 months)
My role was initially to explore the possibilities of applying the graphic style used by Zenly for its emojis to any emojis from another source.
Then, the objective was to validate applicable discoveries, submit these new features, and accompany them to their production.

I was able to experiment with various methods to apply this style transfer, even though the difficulty lay in the fact that there were very few example images. Experiments were mainly conducted using Pytorch and Google Colaboratory notebooks, as well as VMs on GCP (Google Compute Engine).
Nevertheless, good results allowed the validation of the mission's exploration phase and the identification of interesting features for their production.
Data Science Deep Learning Computer Vision Pytorch TensorFlow Python Google Cloud Docker Agile Methodology
Zenly
Infrastructure system migration
INTERNET OF THINGS (IOT)
January 2022 - April 2022 (3 months)
Paris, France
The production environment being driven by an infrastructure-as-code system that produces YAML files to configure production, my role was to update the definition files to make them compatible with the new system version. I also had to ensure that the newly produced YAMLs were compliant.

My mission also involved jointly developing the new version of the infrastructure system, adding validation tools and tests.
Finally, tools were developed for the infrastructure team and developers to update or create new infrastructure description files (in Typescript).

I also developed a tool that is a web application (Express/Vue.js) to visualize the infrastructure's status, with each component and its information and dependencies.
Typescript Node.js Vue.js Docker Back-End Development Google Cloud Mocha Chai DevOps Kubernetes Agile Methodology