Projects & Case Studies

Professional Projects

Below are selected projects demonstrating my expertise in AI/ML, Data Engineering, and Cloud Architecture. While specific client details are confidential, these case studies illustrate the types of challenges I solve.

Real-Time Energy Optimization Platform

Client: Major energy company (GreenFlex - Total Énergie)

Challenge: Optimize energy consumption across 500+ industrial sites in real-time

Solution:

  • Designed streaming data pipeline processing 10M+ sensor readings/day

  • Built ML models for predictive consumption analysis

  • Implemented anomaly detection for equipment failures

  • Created real-time dashboard for operations teams

Technologies: Apache Kafka, Spark Streaming, Python, AWS (Kinesis, Lambda, S3), TimescaleDB

Impact:

  • 25% reduction in energy costs

  • 40% faster anomaly detection

  • €2M+ annual savings

Predictive Maintenance for Rail Infrastructure

Client: SNCF Réseau (French National Railway)

Challenge: Predict equipment failures before they cause service disruptions

Solution:

  • Developed ML pipeline analyzing historical maintenance data

  • Built ensemble models (Random Forest, XGBoost, LSTM)

  • Integrated with existing CMMS systems

  • Automated alerting system for maintenance teams

Technologies: Python, scikit-learn, TensorFlow, PostgreSQL, Airflow, Docker

Impact:

  • 35% reduction in unplanned maintenance

  • 15% increase in equipment uptime

  • Prevented 50+ service disruptions in first year

LLM-Based Document Classification System

Client: Luxury goods manufacturer (TAG Heuer - LVMH)

Challenge: Automatically classify and route 10,000+ customer service documents monthly

Solution:

  • Fine-tuned transformer models (BERT, RoBERTa) on domain-specific data

  • Built API service for real-time classification

  • Implemented active learning pipeline for continuous improvement

  • Created multilingual support (French, English, German, Chinese)

Technologies: Hugging Face Transformers, PyTorch, FastAPI, Docker, Kubernetes, MLflow

Impact:

  • 92% classification accuracy

  • 80% reduction in manual routing time

  • Processing time reduced from 2 days to 5 minutes

Cloud Data Platform Migration

Client: Multiple enterprise clients

Challenge: Migrate on-premise data infrastructure to cloud-native architecture

Solution:

  • Designed cloud-native architecture (AWS/GCP)

  • Migrated ETL pipelines to Spark on Databricks

  • Implemented data lake using Delta Lake

  • Set up CI/CD pipelines for data workflows

  • Trained client teams on new platform

Technologies: AWS (S3, EMR, Glue), Databricks, Delta Lake, Terraform, GitLab CI

Impact:

  • 40% infrastructure cost reduction

  • 10x faster query performance

  • 60% reduction in pipeline development time

Computer Vision for Drone Inspection

Client: Drone Volt Group

Challenge: Automate infrastructure inspection using drone footage

Solution:

  • Built CNN models for defect detection in images/video

  • Developed real-time inference pipeline

  • Created web interface for inspection reports

  • Trained models on custom-labeled dataset (50K+ images)

Technologies: PyTorch, OpenCV, FastAPI, React, AWS SageMaker

Impact:

  • 70% faster inspection process

  • 95% defect detection accuracy

  • Reduced human error by 85%

Time Series Forecasting for Industrial Production

Challenge: Forecast production volumes and resource requirements

Solution:

  • Implemented Prophet, LSTM, and Transformer-based forecasting models

  • Built automated retraining pipeline

  • Created interactive forecasting dashboard

  • Integrated with ERP systems

Technologies: Prophet, TensorFlow, Plotly Dash, Airflow, PostgreSQL

Impact:

  • 15% improvement in forecast accuracy

  • Better resource allocation planning

  • €500K+ annual savings in inventory costs

Open Source & Personal Projects

Data Science Blog & Tutorials

Medium: Technical articles on ML & Data Engineering

Topics covered:

  • Apache Spark optimization techniques

  • MLOps best practices

  • Databricks workflow patterns

  • NLP with transformers

Teaching Materials

GitHub: Code repositories for courses

  • Spark training materials (50+ consultants trained)

  • Data Science project templates

  • ML pipeline examples

Areas of Expertise

ML/AI

Supervised/Unsupervised Learning, NLP, Computer Vision, Time Series, LLMs, Anomaly Detection

Data Engineering

Spark, Kafka, Airflow, Delta Lake, Data Lakes, Streaming, ETL/ELT

Cloud & MLOps

AWS, GCP, Azure, Databricks, MLflow, Kubernetes, CI/CD

Industries

Energy, Transportation, Luxury Goods, Finance, Aerospace

Interested in Working Together?

I’m available for:

  • Consulting on AI/ML projects

  • Architecture design for data platforms

  • Technical training for your teams

  • Speaking engagements at conferences/meetups

Contact: hadjyoucef.amine@gmail.com

Learn more about my background