nadaboulaares

Nada Boulares

Data Scientist (AI, ML, DL, AWS)

Based in Tunisia | Open to Remote Opportunities

Summary

Experienced data scientist specializing in machine learning, deep learning, and data visualization. Proficient in Python programming and databases such as MySQL and MongoDB. Strong track record of developing and deploying models, integrating recommendation systems, and collaborating on impactful projects. Successfully delivered projects with measurable outcomes and enhanced user experiences.

Skills

Experience

ML Engineer, Omdena, New York, United States

April 2022 - Present

Co-led the modeling task of the Minktek project. Collaborated with a multidisciplinary team to develop advanced models and algorithms, resulting in improved accuracy and personalized treatment plans.
Co-led the deployment task of the Belyntic project. Utilized Streamlit and FastAPI frameworks to create user-friendly interfaces and ensure seamless integration. Conducted comprehensive data collection leveraging web scraping techniques with BeautifulSoup and Selenium.
Contributed to the BigCodeGen project, predicting short-term traffic flow congestion on urban motorway networks. Labeled datasets using Labelbox and assisted in creating an object detection model using YOLOv7.
Participated in the Dr CADx project, focused on reducing the risk of misdiagnosis of chest X-rays with deep learning. Involved in data collection from various resources and contributed to exploratory data analysis.
Contributed to the Sysconverge project, optimizing energy consumption for the real estate market using computer vision. Collected data from multiple sources, participated in preprocessing tasks and modeling, and labeled datasets using MakeSense.AI and Label Studio. Assisted in creating an object detection model for doors, windows, and plumbing using YOLOv5.
Collaborated on the MangoSciences project, utilizing AI to support the affordability of oncology immunotherapy treatments. Collected data from various sources, worked on modeling tasks, and assisted in creating survival models. Developed a Cox-time model to predict progression-free survival (PFS).

Data Science Intern, Dance4Healing, California, United States

August 2022 - Present

Integrated a recommendation system on AWS using AWS Personalize, S3, and EC2, enhancing the user experience and driving personalized recommendations.
Applied matrix factorization techniques to solve the cold start problem, improving the system’s ability to make accurate recommendations for new users or items.
Assisted in building the project architecture, utilizing AWS RDS and AWS DynamoDB for efficient data storage and retrieval.
Developed a recommendation system using machine learning techniques such as K-Means and Scipy. Explored the Dance4Healing database in MongoDB, downloaded relevant datasets, and performed data generation. Utilized data visualization libraries like Seaborn and Matplotlib. Created a machine learning model to recommend dance session videos based on factors such as emotion, duration, and energy level. Deployed the model using Flask and contributed to version control on the organization’s GitHub repository.
Contributed to the translation of videos from English to Chinese by leveraging AWS services such as S3, AWS Transcribe, AWS Translate, and AWS Elastic Transcoder, enabling multilingual accessibility.
Collaborated on creating a user dashboard, utilizing tools such as Figma and Python libraries like Matplotlib and Seaborn. Designed a new desktop interface to enhance user interaction and visual representation.

Data Science and Business Analytics Intern, Sparks Foundation, Singapore

June 2022 - July 2022

Conducted data analysis and visualization to derive meaningful insights.
Employed decision tree algorithms for prediction tasks.
Developed a face mask detection system using OpenCV, Keras, and TensorFlow.

Data Science Intern, REGIM Lab, Sfax, Tunisia

February 2021 - September 2021

Deployed the developed model in a mobile application using Flutter and Django, ensuring seamless integration and accessibility.
Successfully developed a deep learning model for facial recognition using Insightface and ArcFace algorithms, utilizing Nvidia, CUDA, TensorFlow, and Keras frameworks.
Conducted a comprehensive comparison of pre-trained models in the Mxnet framework to evaluate their performance and suitability for specific tasks.
Prepared face detection models, including MTCNN, RetinaFace, and Antelope, for accurate facial analysis and identification.
Collected and meticulously prepared data for use with the Mxnet library.