Freelancer profile translated to English.

Description

With 3.5 years of experience in Python, I am motivated to undertake freelance missions in AI systems. My goal is to provide an optimal experience for users and clients across various sectors (retail, banking, insurance, luxury).

Languages

English
Native or bilingual
French
Fluent

Workplace preferences

Can work on-site

Paris (up to 50km)

LIAS
Decision Model Engineer
October 2021 - February 2025 (3 years and 4 months)
- Design and development of optimization and prediction models;
- Translate business specifications into model and technical specifications;
• Develop optimization models in Python with constraints and objective function;
• Contribute to accelerating solver computation time by identifying critical patterns affecting the optimization algorithm, which improved computation time by up to 177 times.
• Perform tests to validate the obtained model by retrieving data in Pandas and JSON format.
Operations Research 3.5 years of experience in Python
Projet personel
RAG System Engineering
October 2025 - November 2025 (1 month)
RAG Chatbot with FAISS and LLMs: Development of a retrieval-augmented generation (RAG) system for structured data:

- Decomposition of each data row into independent chunks for semantic embedding creation;
- Generation of embeddings using a free, offline-capable model (BAAI/bge-m3);
• Construction of a FAISS index to retrieve the most relevant chunks from a query.
• Integration of an LLM (local Ollama or OpenAI API) to generate contextualized natural language responses.
• Design of the pipeline with a clear separation between offline ingestion and online retrieval, ensuring reproducibility and production readiness.
For more details, please refer to the README file available on Github:
https://github.com/nourkarnib/rag-chatbot-system.
RAG LLM Langchain Chatbot Faiss
Projet personnel
Fine-tuning an LLM (GPT-2)
August 2025 - October 2025 (2 months)
- Generative AI for product review analysis (Amazon reviews): can be used as a language model in a complex AI system, such as a chatbot or a multimodal model.
- Preprocessing of real-world noisy input data;
- Fine-tuning GPT-2 for domain-specific analysis, reducing perplexity and improving model performance;
• Management of large-scale version control for checkpoints and model artifacts.
• Deploying the model on Hugging Face Hub, demonstrating end-to-end solution delivery, available here: https://huggingface.co/spaces/nourkarnib/fine-tuned-gpt2.
• Structured engineering solutions for improved model input structure, reflecting practical problem-solving in complex systems.
For more details, please refer to the README file available on Github: https://github.com/nourkarnib/Gen-AI-Opinion-about-products-based-on-AWS-data.
LLM Python OpenAI API