I am an Software Engineer specializing in Data Science, React.Js, and Java. My expertise covers:
- RAG (Retrieval-Augmented Generation)
- LLMs (GPT, Azure OpenAI, LLaMA)
- Microservice
- Microfrontend
I excel at designing and deploying end-to-end AI solutions (chatbots, real-time data sync, continuous training with MLflow). I’m driven by a passion for cutting-edge AI research and a commitment to continuous learning.
- Location: Sfax, Tunisia
- Email: mohamedamine.macherki@ieee.org
- Phone: +216 56 58 60 61
- LinkedIn: Mohamed Amine MACHERKI
- GitHub: Mohamed Amine MACHERKI
- LLMs: GPT, Azure OpenAI, LLaMA
- Fine-Tuning & Prompt Engineering
- RAG: Retrieval-Augmented Generation
- Hugging Face: Transformers, Pipelines
- Supervised & Unsupervised ML (Regression, Classification, Clustering)
- Computer Vision, NLP, Time Series Analysis
- Transfer Learning
- MLflow for experiment tracking and continuous retraining
- Databases: PostgreSQL, PGVector (Vector DB), InfluxDB, Redis
- DevOps Tools: Docker, Kafka, Selenium, ETL (SSIS), Grafana, Power BI
- Languages: Python, SQL, Java, C
- Frameworks: Flask, FastAPI
-
Generative AI Chatbots:
Built a conversational chatbot on AWS using OpenAI GPT-3.5 Turbo and PGVector for real-time retrieval, synchronized via Kafka events.
-
LLM-based Virtual Assistants:
Developed domain-specific assistants (Finance, HR, Product Owner) integrated with Jira using RAG Agents to automate deliverables (PDFs, roadmaps).
-
Energy Optimization Platform:
Predicted and optimized energy production/consumption with advanced ML algorithms; real-time monitoring via Grafana/Power BI.
-
Audio Leak Detection:
Created a commercialized deep learning model (97% accuracy) for industrial leak detection (Wedetect).
-
ECG Classification:
Implemented MobileNet-based ECG classification (95% accuracy) for medical decision support.
Explore more on my GitHub profile »
