FlowCV Logo
Work Experience
AI Engineer, Deloitte
Jan 2024 – present | Cairo, Egypt
  • •LLM Agent Development and Deployment for a medical company
  • •Led the architecture and system design of an LLM-powered solution for extracting and summarizing Power BI data using Google ADK and Vertex AI. Designed a layered, pipeline-based architecture that separates data parsing, summarization, and deployment into modular components. Applied Modular Design Principles like separation of concerns, single responsibility, and loose coupling to ensure reusability and maintainability. Implemented strategy-based agent orchestration and deployed the system on Google Cloud Run using containerization.
  • •Data Migration and Validation for a Leading Oil Company
  • •Implemented data migration processes using DBT integrated with Snowflake and performed data validation using SQL to ensure accuracy.
  • •Data Mining Application for an Insurance Company
  • •Led the design and development of the backend of a data mining application tailored for the insurance sector, incorporating a MongoDB database using fast api.
  • •Integrated bootstrapping actuarial calculations and analytic tools to extract insights from large datasets.
  • • AI Institute Researcher
  • •Designed the vector database on milvus for RAG
  • •Actively contributed to the development of advanced transcription systems leveraging state-of-the-art speach to text local models (NeMo and WhisperX), resulting in accurate speach-to-text capabilities.
  • •Collaborated in the design and implementation of LLM agents using Langchain and Langgraph.
    • Data Scientist, MF Strategy , Part-time
    2023 – 2024 | Lausanne, Switzerland
  • •EWS for a major Egyptian national bank
  • •Developed a behavioral Early Warning System (EWS) model to predict the probability of clients exceeding a payment delay threshold utilizing PySpark for scalable data processing and model development.
    • Data Scientist, Al Tadamun Microfinance Foundation
    Jan 2023 – Jan 2024 | Cairo, Egypt
  • •Analyzed raw data to extract KPIs and create interactive PowerBI dashboards for strategic decision-making
  • •Collaborated with IFC to develop a credit scoring system with 80% accuracy for better risk assessment.
  • •Conducted R&D to optimize client loan amounts by integrating classical machine learning, financial methods,clustering techniques, and reinforcement learning for improved risk management and profit optimization.
  • •Developed an automated ID card extraction system with data collection and annotation, Fast R-CNN model training, image enhancement using OpenCV, OCR text extraction, and a deployable API for Flutter integration.
  • •Developed an ID verification system using facial recognition , employing the VGGFace2 model for facial identification and conducting a face-to-face comparison between two images, ensuring secure identity verification.
  • •Forecasted the probable sales for the next quatres to be used to make quatre plans using multiple models(ARIMA ,RNN).
    • Education
    Postgraduate Diploma of Ai and machine learning, Institute of information Technology (ITI)

    9-month program powered by École Pour l'Informatique et les Techniques Avancées

    Bachelor of Biomedical Engineering, Helwan University

    Project: computer vision for breast cancer detection

    Project grade: excellent

    1 / 2
    Skills
    Programming Language

    Python, SQL, R, Linux shell scripting, C++

    Architecture & Design:

    Solution Architecture, Software Design Patterns (Factory, Pipeline, Strategy), Modular System Design, Scalable Cloud Architecture.

    Mathematics and Statistical analysis
    • •Hypothesis Testing , Confidence Intervals and A/B testing
    • •Descriptive and Inferential Statistics
    • •Linear Algebra (matrix operations, eigenvectors, SVD)
    • •Calculus (gradients, optimization in ML models)
    • •Cost Functions and Loss Metrics (MSE, Cross-Entropy, LogLoss)
    Machine Learning and Deep Learning
    • •Supervised Learning: Regression, Classification (Logistic Regression, Random Forest, XGBoost, SVM, etc.)
    • •Unsupervised Learning: Dimensionality Reduction (PCA, TSNE), Clustering (KMeans, DBSCAN, etc.)
    • •Computer Vision : Object Detection and Image Classification using CNN variants (e.g., VGG, ResNet, yolo), Image Preprocessing and Enhancement (OpenCV, PIL)
    • •Time-Series Forecasting: Trend, Seasonality, and Noise Decomposition Forecast Modeling using ARIMA, RNN, and LSTM
    • •NLP: Text preprocessing, Sentiment Analysis, NER, Text Classification, Summarization, Keyword Extraction
    • •Model Optimization: Feature engineering, cross-validation, hyperparameter tuning
    • •Machine Learning Frameworks: scikit-learn, TensorFlow, Pytorch, MLlib
    Deployment
    • •Deployment Frameworks: FastAPI, Flask
    • •Containerization: Docker
    Cloud

    Azure, AWS, Databricks, Dataiku

    Database and Data Engineering
    • •Databases: PostgreSQL, Snowflake, MongoDB
    • •ETL & Data Pipelines: Design and deployment of scalable ETL workflows using Airflow, Azure Data Factory (ADF) , and DBT
    • •Data Modeling: Relational and document-based schema design, normalization, incremental loading strategies
    • •Data Migration & Validation: End-to-end data migration using DBT; automated validation via SQL queries and conditional logic
    • •Data Processing: PySpark, Pandas
    • •Query Optimization: SQL tuning for performance, partitioning, indexing strategies
    • •Data Governance: Handling of schema lifecycle (e.g., Alembic + SQLAlchemy), data quality checks, versioned deployments
    Data Visualization and Dashboards
    • •Business Intelligence: Power BI, Tableau
    • •Python Visualization: Matplotlib, Seaborn, Plotly Dash
    • •Skills: Interactive Dashboard Design, KPI Tracking, Visual Storytelling
    LLM Agents:
    • •LLM Theory: Transformer architecture, self-attention, multi-head attention, encoder-decoder models, positional encoding, fine-tuning methods
    • •LLM Agent Frameworks: LangChain, LangGraph, Google ADK
    • •Prompt Engineering: Structured and optimized prompts for summarization, classification, and extraction tasks
    • •Retrieval-Augmented Generation (RAG): Integrated Milvus vector databases for context-aware information retrieval
    • •Speech-to-Text: WhisperX, NeMo
    • •LLM Agent Design: Multi-agent workflows, task orchestration,
    Certificates
    2 / 2