About

Professional experience

• Senior Data Scientist, Data Science Team, AXA France

  June 2021 - Present

  • Developed text extraction AI Systems using computer vision to read documents (Driver's license, National ID card, etc.). The models processed over 6M documents in batches using PySpark Jobs

  • Developed a range of reusable python packages for document processing (OCR, object detection, etc.) The packages are designed using the strategy pattern to quickly experiment with different algorithms / models.

  • Automated the indexing of 70% of AXA France's incoming documents by developing a retrainable document classification AI System (exposed as an API).

  • Created a YAML-based templating system to streamline the creation of Azure ML pipelines.

  • Built reusable Azure ML Pipelines (Continuous Training) to automatically retrain AI Systems and publish to Model Registry.

  • Designed Azure DevOps CI/CD/CT pipeline templates to be used in the Team's projects.

  • Led the hiring process, onboarding new team members and facilitating efficient knowledge transfer.

  • Coordinated the annotation process with annotators and SMEs (subject matter experts) to ensure consistent datasets across various projects.

  • Led AXA's 'Python COP' initiative, organizing a series of monthly programming talks and events on topics such as testing, web scraping, packaging, code quality, etc.

  • Teached a series of training sessions/courses on Python, Data Manipulation (pandas & pyspark), Software Engineering best practices for over 30 AXA collaborators.

• Graduate Data Scientist, Data, Fraud, Waste & Abuse Team, AXA France

  May 2020 - June 2021

  • Engineered an advanced AI-driven Fraud, Waste, and Abuse Detection Tool that inspects different scopes, including Claims, Health Providers, Beneficiaries, and more. This innovative tool played a pivotal role in achieving savings exceeding 150k euros in 2022.

  • Designed and trained various fraud detection models for fraudulent patterns as part of the FWA detection tool.

• Data Scientist, Shift Technology

  April 2019 - March 2020

  • Applied Natural Language Processing (NLP) techniques to cluster Claim managers' notes, leveraging advanced visualization methods to uncover and identify emerging fraud patterns.

  • Built new AI Models for Shift's fraud detection solution & Deployed their solution for a new customer.

• Research Internship, U2IS - ENSTA Paris

  May 2018 - August 2018

  • Created a synthetic image dataset by building a simulation of an agent in an in-house environment and using Raytracing to render realistic images.

  • Studied Catastrophic Forgetting in Incremental Learning of Principal Components of generated Images

Projects

• mlflower: Lightweight orchestration tool for mlflow projects

  November 2023 - Present

  Tool to extend mlflow projects with inbuilt multi step workflow orchestration

• pysira: publish resumes in different formats

  February 2023 - Present

  CLI tool to export 'jsonresume' files to different formats (HTML, TeX, PDF, etc.) and languages.

  • Created a CI/CD pipeline (Github Actions) to publish resumes in different themes, formats and languages

  • This resume was created using pysira: hmiladhia.github.io, hmiladhia.github.io/cv.pdf, hmiladhia.github.io/fr/kendall or hmiladhia.github.io/resume.pdf

• nbmanips: Split, merge and convert IPython Notebooks

  May 2021 - Present

  An open-source package containing a collection of tools to manipulate Notebooks via a python or CLI.

  • Developed a python package/CLI to manipulate (Split, Merge, ...) Jupyter Notebooks

  • Used CI/CD piplines (Github Actions) to Lint, Test and publish the package to Pypi

• Spot-Language: Programming Language Detection

  January 2020

  Spot-Language is a classification model to detect the language used in code snippets.

  • Assembled a training dataset using public repositories on GitHub.

  • Built an ML experiment to train an NLP model for language classification.

  • Improved model interpretability using Lime.

  • Deployed a demo Web Application of the ML model using flask.

• Dmail

  April 2020 - May 2020

  A python package to write & send emails in markdown format

  • Developed a python package to write and send emails in markdown format

  • Used CI/CD piplines (Github Actions) to Lint, Test and publish the package to Pypi

• DroNet

  April 2019 - May 2019

  DroNet for the forest environment

  • Built a dataset from different sensors and a camera of a terrestrial mobile robot placed in a forest environment.

  • Fine-tuned the DroNet neural network on the new dataset and tested the model in a Bebop drone

• ENIT Robots competition

  April 2017

  Participation in ENIT Robots competition as team leader. Obtained 3rd place.

• ENSIT Robots competition

  December 2017

  Development in python of a trajectory recognition algorithm for a line following robot equipped with Raspberry Pi 3 and a camera.

Skills

Education

Certificates