Name: AbdelKarim ELJANDOUBI
Type: User
Bio: I am a Data Scientist, graduated with engineering degree from ENSTA Paris. I also have a master's degree in mathematics, vision, machine learning from ENS Paris
Location: Massy, Île-de-France.
Blog: https://www.linkedin.com/in/abdelkarim-eljandoubi/
AbdelKarim ELJANDOUBI's Projects
BANK PROFIT NAVIGATOR : Extraction of crucial financial metrics from financial reports.
A web app that allows you to select a subject and then change its background, OR keep the background and change the subject.
Airflow data pipeline
Build an ETL pipeline for a database hosted on AWS Redshift.
Using AWS Glue, AWS S3, Python, and Spark, create or generate Python scripts to build a lakehouse solution in AWS
Implementation of the Breakout Strategy
In this project, I've used AWS Sagemaker to build an image classification model that can tell bicycles apart from motorcycles. And I've deployed the model, using AWS Lambda functions to build supporting services, and AWS Step Functions to compose the model and services into an event-driven application.
Build an ML Pipeline for Short-Term Rental Prices in NYC
Homemade Copilot: Fine-tune LLM through LoftQ initialization and QLoRA-style training for code generation.
Simple CRM backend written in Go
An implementation of DDPG agent to solve a Unity environment like Reacher and Crawler.
Deploying a ML Model to Cloud Application Platform with FastAPI
Project aim is to build a Natural Language Processing (NLP) model to categorize messages on a real time basis.
Speech Recognizer basic models.
Use AWS Sagemaker to train a pretrained model that can perform image classification by using the Sagemaker profiling, debugger, hyperparameter tuning and other good ML engineering practices.
An implementation of Deep Q-Learning Network for solving a Unity environment that can navigate and collect bananas in a large, square world.
ML Model Scoring and Monitoring
Generate new faces using Generative Adversarial Networks (GANs).
In that project, I combined my knowledge of computer vision techniques and deep learning architectures to build a facial keypoint detection system. It was designed to take in any image with faces and predict the location of 68 distinguishing keypoints on each face.
Create an ML pipeline for Genre Classification using MLflow.
Fine-tune the Vision Transformer (ViT) using LoRA and Optuna for hyperparameter search.
Utilize Reinforcement Learning (RL) to address challenges in the platform environment characterized by a hybrid action space.
Uses a CNN Encoder and a RNN Decoder to generate captions for input images.
a CNN-powered app to automatically predict the location of the image based on any landmarks depicted in the image.
An implementation of MADDPG multi-agent to solve a Unity environment like Tennis and Soccer.
Deploy an app to multi-cloud