Serine Daouk

LinkedIn   |  Email: serine_md@hotmail.com  |  Phone: +9613829424  |   Github  |  Location: Beirut, Lebanon 

Skills
Work Experience
  • Machine Learning and MLOps Engineer
      CoordinatesME - Dubai (remote)
      February 2024 - Present
    • Lead and maintain the end-to-end development of a real-time classification system, processing 24/7 signal streams with less than 2s latency per request under concurrent loads. Built scalable service-based backend using FastAPI, RabbitMQ, BentoML, MLflow, Docker, Qdrant, Grafana, Prometheus, Evidently, and XGBoost.
    • Fine-tune and deploy a customized open-source LLM using Ollama, PyTorch and Unsloth on a local NVIDIA GPU server to maximize inference performance and resource utilization.

  • Backend and Machine Learning Engineer
      AITECHOM - Muscat, Oman (remote)
      June 2022 - Nov 2023
    • Built the main processing backend component of a cyberattack detection system by implementing a PySpark distributed cluster for processing a high volume of incoming network traffic.
    • Managed and integrated the Machine Learning components with MongoDB as the underlying data store for insights, ensuring efficient storage and retrieval of model-generated data.
    • Implemented PySpark Streaming with Apache Kafka data lake source and MongoDB database and developed, implemented, and tuned machine learning models for real-time cyberattack detection (e.g. ransomware, SMB brute force, port scanning, etc) using Scikit-Learn.
    • Designed and implemented the automated retraining pipeline and MLOps workflow for the behavioral-based ML model for real-time drift detection with DVC, MLFlow, and River-ML.

  • Software and Machine Learning Engineer
      Aghive - Beirut, Lebanon
      March 2022 - June 2022
    • Designed and implemented a data processing pipeline (ETL) to effectively process satellite images obtained from Google Earth Engine for object detection.
    • Collaborated on enhancing satellite image resolution using an ESRGAN model, aimed at improving the quality of imagery for more accurate detection tasks.
    • Leveraged the Meta Detectron2 library to orchestrate the training of a Mask R-CNN model on annotated agricultural field data, enabling precise instance segmentation for agriculture-related applications.
    Education
    Bachelor of Engineering in Computer and Communication Engineering
    American University of Beirut
    August 2017 - August 2021
    Course Focus: Algorithms & Data structures, Software Design & Architecture, Machine Learning and Neural Networks, Computer Networks, Computer Architecture.
    Achievements & Certificates
    • AWS Certified Machine Learning - Specialty.
    • AWS Certified Cloud Practitioner.
    • Certificate of Completion in Deep Learning: Advanced Computer Vision and Object Detection course.
    • Certificate of Achievement in a Pre-Engineering Senior Robotics track.
    • Certificate of Completion in the Hands-On Machine Learning with scikit-learn and Python course.
    Personal Projects
    Personal Resume Static Website
    • Deployed a static website to a custom domain with HTTPS using S3, CloudFront, Route53, and Certificate Manager.
    • Implemented a dynamic back-end page view counter using Lambda (Python), DynamoDB, API Gateway, JavaScript, and Fetch API.
    • Set up a robust CI/CD pipeline for redeploying front-end and back-end website components, leveraging GitHub Actions integrated with AWS.
    Pyspark Structure Streaming Template Repository
    • Created a Guthub Repository template showcasing how to set up Apache Pyspark Structured Streaming.
    • Demonstrated how to connect data sources and sinks like Apache Kafka and MongoDB to a Pyspark Streaming Engine.
    Smart Virtual Desktop Assistant
    • Designed a cross-platform desktop assistant using python and machine learning engines for the backend.
    • Handled the voice-to-text processing part to prepare data for the models.
    Image Classification Mobile Application
    • Trained a model to classify between cat & dog images.
    • Handled the voice-to-text processing part to prepare data for the models.
    • Saved the model using TFlite and integrated it into a flutter based cross-platform mobile application.
    Hobbies
    • Tennis, ski, running
    • Reading fiction & non-fiction books.
    • Playing sudoku or chess.