Mohamed Khalil ZRELLY

Generative AI Engineer Apprentice — Crédit Agricole Personal Finance & Mobility

Passionate about Generative AI and Agentic Systems, I build on solid machine learning foundations to engineer intelligent agents that reason, retrieve, and deliver real-world impact.

Who am I?

I'm a Computer Science Engineering Student at Sup Galilée, Sorbonne Paris Nord University, and a Generative AI Engineer Apprentice at Crédit Agricole, focused on agentic AI systems, with a solid foundation in machine learning, deep learning, and NLP.

Beyond technology, I was involved with Enactus and AIESEC in social entrepreneurship and international exchange, which reflects the same curiosity, leadership, and dedication I consistently bring to every complex engineering challenge I encounter along the way.

Where have I worked?

  • Generative AI Engineer Apprentice — Crédit Agricole Personal Finance & Mobility

    09/2025 — Present · 1 year 2 months · Massy, France

    • Developed a declarative, config-driven microservice for workflow management and orchestration, making it easy to manage and execute complex AI workflows at scale.
    • Built an automated workflow evaluation engine leveraging LLM-as-a-judge to assess outputs against custom metrics, with a dashboard to analyse workflow behaviour.

    Agentic AI Workflows, LangGraph, RAG/MCP, Microservices, Python/FastAPI, Amazon Bedrock, Amazon DynamoDB, Workflow Evaluation

  • Natural Language Processing Intern — VERMEG for Banking & Insurance Software

    02/2024 — 06/2024 · 5 months · Tunis, Tunisia

    • Implemented a system that automates email classification and information extraction from attachments, using fine-tuned BERT and OCR to improve accuracy and optimise efficiency.
    • Developed a web application for BERT Fine-Tuning, email management, and real-time data analysis and visualisation, enabling model experimentation and performance monitoring.

    PyTorch, Hugging Face Transformers, Fine-Tuning, Sentence-Transformers, Python/Flask, MongoDB, PaddleOCR, React/Redux, Bootstrap

  • Machine Learning Intern — Sofiatech

    06/2023 — 08/2023 · 3 months · Ariana, Tunisia

    • Evaluated multiple reinforcement learning algorithms from Stable Baselines 3 across various Gymnasium environments to assess agent performance and training efficiency.
    • Built a 2D game in Pygame where a Deep Q-Learning agent learns to navigate a dynamic environment, applying the best-performing algorithm from the benchmarking phase.

    Reinforcement Learning, Deep Q-Learning, stable-baselines3, Gymnasium, TensorFlow/Keras, Python/Pygame, NumPy, Pandas, Matplotlib

What have I built?

  • AeroNU

    06/2026 — Present

    Designing a multimodal travel guide that generates personalised travel itineraries, curated destination recommendations, and immersive visual content, covering the full journey from planning to discovery.

    Multimodal LLM, Advanced RAG, Vector Search, Knowledge Graph, Curation, Visual Generation, Text-to-Image Generation

  • PlanYV

    06/2026 — Present

    Designing a generative project management platform that produces complete project structures, tasks, schedules, and assignments, pairing team members to each task based on their skills and availability.

    Agentic Workflows, LLM Planning, Task Decomposition, Scheduling, Resource Allocation, Skill Matching, Delegation

  • TwyXer

    06/2026 — Present

    Designing an intelligent multi-agent system for automated social listening on Twitter (now X), aimed at identifying emerging trends in real time and generating actionable marketing campaign ideas at scale.

    Multi-Agent Systems, Trend Detection, Python/FastAPI, LangGraph, MCP/A2A, GPT-OSS, Qdrant, Vue.js, Docker, X API

  • CVance

    06/2025 — 08/2025

    Created an intelligent CV recommendation system based on job description using Retrieval-Augmented Generation (RAG), to optimise the selection and automatic ranking of the most relevant profiles.

    LangChain, Advanced RAG, Sentence-Transformers, Milvus, Re-Ranking, Qwen3-VL, Python/Streamlit, Prompt Engineering

    https://github.com/mkirell/CVance

  • VERMail

    02/2024 — 06/2024

    Developed an end-to-end email intelligence platform combining fine-tuned BERT for multi-class classification and PaddleOCR for accurate attachment parsing, with a React dashboard for real-time monitoring.

    PyTorch, BERT Fine-Tuning, PaddleOCR, Python/Flask, MongoDB, REST API, JWT Authentication, React/Redux, Chart.js

    https://github.com/mkirell/VERMail

  • Slythera

    06/2023 — 08/2023

    Developed a 2D grid game in Pygame featuring an AI agent trained from scratch with Deep Q-Learning to autonomously navigate obstacles and maximise its score in a dynamic, progressively challenging environment.

    Reinforcement Learning, Gymnasium, Experience Replay, TensorFlow/Keras, Python/Pygame, NumPy, Pandas, Matplotlib

    https://github.com/mkirell/Slythera

What can I do?

  • Programming Languages (PL)

    Python, TypeScript, JavaScript, Java, C/C++, PHP, PL/SQL

  • Data Engineering & Databases

    PostgreSQL, pgvector, MongoDB, PySpark, Airflow, Redis

  • Machine Learning & MLOps

    PyTorch, TensorFlow, Sklearn, XGBoost, Amazon SageMaker, MLflow

  • LLM Fine-Tuning & Evaluation

    Transformers, PEFT, LoRA/QLoRA, TRL/DPO, DeepEval, Langfuse

  • Generative AI & Agentic Systems

    LangChain, LangGraph, RAG/MCP, Amazon Bedrock, Qdrant, Milvus

  • Full-Stack Web Development

    FastAPI, Pydantic, GraphQL, NestJS, React, Vue.js, Redux

  • Cloud Platform & AWS Services

    Amazon ECS/Fargate, AWS Lambda, Amazon S3, DynamoDB, Cognito

  • DevOps, CI/CD & Observability

    Docker, Terraform, GitHub Actions, Grafana, OpenTelemetry

What did I study?

Degrees

  • Engineering Degree in Computer Science, Sup Galilée – Université Sorbonne Paris Nord (2024 — Present)
  • Bachelor's Degree in Computer Science, ISI Ariana – University of Tunis El Manar (2021 — 2024) — Very good
  • Baccalaureate Degree in Mathematics, Ibn Rachik Ezzahra High School (2020 — 2021) — Good

Certifications

  • Build Multimodal Generative AI Applications — IBM (Coursera) (2026-07)
  • Generative AI Advanced Fine-Tuning for LLMs — IBM (Coursera) (2026-07)
  • Spark and Python for Big Data with PySpark — Udemy (2024-03)
  • Natural Language Processing with Python — Udemy (2024-02)
  • Azure AI Fundamentals (AI-900) — Microsoft Azure (2024-01)
  • Azure Data Fundamentals (DP-900) — Microsoft Azure (2024-01)

Languages

  • English — b2
  • French — b2
  • Arabic — c2

What have I achieved?

Volunteering

  • Project Manager — Enactus

    09/2021 — 08/2024

    As Project Manager in Enactus's R&D unit, I led my team to the Ichmilni Race finals, presenting AGRIVEST, a peer-to-peer lending platform promoting digital inclusion for smallholder farmers in rural Tunisia.

  • Team Member — AIESEC

    02/2024 — 07/2024

    As a Team Member in AIESEC's Incoming Global Talent unit, I negotiated partnerships with companies and educational institutions in Tunisia to secure internship opportunities for international exchange participants.

Awards

  • Hult Prize ISI Ariana Second Place — Tunisia
  • Action With Africa Vice Champions — Netherlands
  • Ichmilni Race Enactus Tunisia Finalists — Tunisia
  • Enactus World Cup Vice Champions — Puerto Rico

How to reach me?

Currently working as a Generative AI Engineer Apprentice at Crédit Agricole Personal Finance & Mobility. Open to new opportunities and conversations.