Mars

Senior Data Scientist

CrossTech

As the lead data scientist, I oversee, define and implement the data mining and statistical modelling tools needed to analyze the UK railway network behaviour for millions of trains.

• Big data analysis: Build a custom data pipeline using Prefect and Prefect Cloud fetching, cleaning, and transforming movement data of billions of trains journey into a suitable format for data analysis.

• Multivariate Statistics, Modelling and Machine Learning: Develop mathematical models understanding trains delay propagation though the railway network using a wide range technic from Generalized Linear Model, statistical hypothesis testing to classification and train delay prediction, leveraging the Python data science stack with Pandas, NumPy, Scikit-learn and more.

• High Dimensional Data Analysis: Feature engineering extracting meaningful insights using a wide range of unsupervised learning methods for data dimensionality reduction.

• Complex Data Visualization: Implement comprehensive visualization of complex patterns and phenomena’s using Python visualization library (Seaborn, Matplotlib or Plotly for interactive visualization allowing dynamic interaction with the data).

• API building: Using the Python FastAPI framework, build an API to power a dashboard used by performance analysts in the railway industry.

Lead Data Scientist

CrossTech is a London-based tech startup specializing in building cutting-edge technology products that fundamentally improve: asset management, operational performance, track worker safety, and the quality of passenger information within the rail and bus industries.

As a Senior Data Scientist, I drove the data science part of a high-profile project dedicated to improving the railway network performance using train movement data.

CEA - Commissariat à l'énergie atomique et aux énergies alternatives

Main Data Scientist in a cross-functional team, developing specifics tools analyzing brain development.

• Investigate brain-behavior relationship on high dimensional imaging data using a wide range of statistical and machine learning technics.
• Create and Develop ConPagnon 2.0, a full featured Python library designed to facilitate the statistical analysis of brain imaging data, leveraging Scikit-learn, Nilearn and others Python data analysis library. Using Sphinx, and Sphinx-Gallery, write a complete online documentation for ConPagnon.
• Implement Features Selection and Data reduction pipeline with regularization technics for both high dimensional imaging and clinical data.
• Automate the team data analysis pipelines gluing together neuroimaging software using Python, R, Matlab and Shell scripting.
• Interact with neuroscientists, clinical researchers, and data engineers to diagnoses problems and make technical recommendations.

Data Scientist

Analyzing data and building analysis pipeline for a novel therapy for the Parkinson disease, in a multidisciplinary team led by neurosurgeon Alim-Louis Benabid, MD, PhD.

• Implement a fully automated pipeline from the brain image 3D segmentation of targeted structures to the computation of the predicted delivered drug doses using Monte Carlo simulation. Maintain and update the laboratory documentation.
• Using R, build routines for 2D and 3D data visualization of the collected clinical data.
• Using Python, aggregate and clean Data from previous lab experiment to predict, using linear models, the patient behavior score. Implement Features selection/ranking or Data reduction.
• Redaction in a scientific published article.

Modelling and analyzing data of a new generation of organic CMOS sensors from both real world and simulated datasets.

• Using Python, tune the hyper-parameters of a Support Vector Regression algorithm predicting the output noise of a CMOS sensors.
• Implement Matlab routine generating a simulated input dataset reproducing typical sensors inputs.

I am a Lead Data Scientist, with seven years of experience successfully applying statistics and machine learning methods to various fields, ranging from technology to neurosciences. In addition to my solid mathematical background and coding skills, my open-mindedness allows me to position myself at the intersection of various disciplines, and my sense of pedagogy allow me to convey complex and technical ideas to people collaborating beyond.

My work experience mostly put the emphasis on:
• High Dimensional Data Analysis         
• Multivariate Statistics and Machine Learning
• Coding in Unix Environment: Python, R, Matlab and C
• Data pipeline automation and version control        
• Software Engineering
• Teamwork and Cross-functional communication

Taplio

Dhaif BEKHA's Linkedin Analytics

Dhaif BEKHA

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Dhaif BEKHA's Linkedin Analytics

Dhaif BEKHA

Want to drive more opportunities from LinkedIn?

Dhaif BEKHA's Linkedin Analytics