AI Engineer

Tamer Elkot

Bridging the gap between robust AI models and scalable production systems.

View My Work Download Resume

About Me

Tamer Elkot

I am an AI Engineer specializing in Computer Vision, MLOps, and backend development, with hands-on expertise in designing and deploying end-to-end AI systems—from data pipelines to production inference servers. I work with TensorFlow, PyTorch, YOLOv8, FastAPI, Docker, and PostgreSQL to deliver scalable, production-ready machine learning solutions.

With a strong engineering mindset and a background in Mechatronics Engineering, I have a proven ability to lead cross-functional teams and deliver measurable results. I am passionate about applying AI to solve impactful problems in agriculture, logistics, healthcare, and security, handling the full lifecycle from data engineering to cloud deployment.

Technical Skills

AI & Computer Vision

TensorFlow Keras Scikit-learn YOLOv8 OpenCV Transformers VLMs

Backend Development

FastAPI Django Streamlit PostgreSQL SQLAlchemy REST APIs Celery

MLOps & Infrastructure

Docker Linux (Ubuntu) Git GitHub Actions AWS VectorDB

Languages & Data

Python C SQL Pandas NumPy

Featured Projects

AgriVision

AgriVision

  • Geospatial Pipeline: Engineered a robust 12-channel multispectral data pipeline using Sentinel-2 imagery with advanced radiometric normalization.
  • Deep Learning Architecture: Trained a custom U-Net for semantic segmentation with a hybrid Dice + Binary Cross-Entropy loss.
  • MLOps & Deployment: Deployed via FastAPI for asynchronous processing and Docker for containerization, persisting data in PostgreSQL.
U-Net Sentinel-2 FastAPI Docker PostgreSQL
EmpVision

EmpVision

  • Replaced legacy HOG methods with YOLOv8 and FaceNet, achieving high accuracy at 90° angles and in low-light conditions.
  • Architected a FastAPI and PostgreSQL backend using native ARRAY types to store 512-dimensional face embeddings for vector similarity search.
  • Designed a multi-threaded RTSP pipeline for zero-latency streaming with Passive Video Enrollment to filter blurred frames.
YOLOv8 FaceNet FastAPI Docker PostgreSQL
Employee Turnover Prediction

Employee Turnover Prediction

  • Achieved 99% prediction accuracy using a tuned XGBoost model with SMOTE for class imbalance handling.
  • Dockerized the application and deployed it via Streamlit, making the predictive tool accessible for end-users.
XGBoost SMOTE Streamlit Docker
Depression Rate Survey

Depression Rate Survey (Kaggle Top 8)

  • Built an ML pipeline achieving 92.09% accuracy, ranking Top 8 of 500+ participants using Optuna hyperparameter tuning.
  • Utilized Stratified K-Fold Cross-Validation to ensure robust performance across minority classes.
XGBoost Optuna Scikit-learn

Experience

Jan 2026 – Present

Computer Vision Engineer Intern

Celulla Technology (Remote)

  • Shoplifting Detection: Directed a team to develop a real-time CNN + LSTM & MoViNet system.
  • Water Segmentation: Engineered 'AgriVision' using Sentinel-2 and U-Net for flood monitoring.
  • Teeth Classification: Built an end-to-end diagnostic system using FastAPI and Streamlit.
Jan 2026 – Present

Machine Learning Engineer Intern

Technocolabs Softwares (Remote)

  • Predictive Risk Engine: Building an end-to-end SLA breach predictor for logistics.
  • Data Analysis: Analyzed 200,000+ shipment records; identified primary failure drivers.
  • MLOps Pipeline: Engineered a robust pipeline with data validation for XGBoost modeling.
Jun 2024 – Feb 2025

AI Engineer Trainee

Neurotech (Remote)

  • Mastered Python OOP and implemented data extraction pipelines using BeautifulSoup.
  • Gained foundational knowledge in AI principles and industry applications.

Education & Certifications

Education

Bachelor of Science in Mechatronics Engineering

Higher Institute of Engineering & Technology (Expected: May 2027)

Certifications