About

Hello! I'm Adrien Kouki (aka Kooks)

I have a passion for ML applications and data engineering. I'm currently an ML engineer at L'Oréal, deeply engaged in the Worldwide cosmetics regulations project, which entails agentic AI, US, EU and CN regulatory landscape. A challenging context in which I build tools for automatic analysis and exposure. I work mainly with Python, GCP, SQL, Go, FastAPI, and AWS.

Why have this blog?

Over the years, I've tackled some pretty interesting challenges in ML systems and data infrastructure. This blog is where I share the solutions, gotchas, and lessons learned from those experiences. Think of it as my technical notebook made public.

Writing about complex problems forces me to really understand them, and I've found that other engineers often face similar challenges. So why not share what works (and what doesn't)? Each post is essentially me saying "here's what I learned the hard way, maybe it'll save you some time."

Always happy to discuss different approaches and hear how others have solved similar problems .

My credentials

I hold a Master's degree in Computer Science from IMT Atlantique, where I specialized in Machine Learning and Data Engineering. I also hold a Research Master's in Artificial Intelligence from Université Paris-Saclay. Additionally, I've obtained several certifications:

AWS Certified Machine Learning Associate

AWS Certified Machine Learning Associate - early adopter

• 

  Mar 2025 – Present

  L'Oréal

  Machine Learning Engineer

  • Generative AI Implementation: Designing and implementing advanced prompt engineering systems for automated marketing claims evaluation, reducing manual review time by 80%.
  • Data Pipeline Architecture: Developing web scraping solutions and building robust ETL processes using dbt and Dagster, enabling seamless data integration across data stores.
  • Knowledge Graph Innovation: Conducting research and implementation studies to map complex relationships between product ingredients, claims, and regulatory requirements, creating a unified linked information ecosystem.
• 

  Nov 2024 – Feb 2025

  Qantev

  ML and Data Engineer

  • Generative AI & OCR Integration: I integrated generative AI models (LLMs) and Optical Character Recognition (OCR) technologies to automate the processing and analysis of legal documents, improving data processing efficiency.
  • API Design & Maintenance: I designed and maintained efficient APIs to facilitate access to AI services and models, ensuring their performance and reliability.Acted as an information relay with product, devops and data model teams.
  • Data Pipelines & Quality Control: I ensured efficient data pipeline integration with rigorous quality controls to guarantee data integrity and efficient information flow within the system.
• 

  Mar 2024 – Oct 2024

  Pelico

  Lead Data Engineer

  • Led a team of 4 data engineers in the development and implementation of supply chain solutions for manufacturing clients in the aircraft, jewelry, and automotive industries. ETL with Apache Airflow on Azure/OVH.
  • Leveraged expertise in modern data engineering tools, including Apache Airflow, Kubernetes, DBT Core, Terraform, and Ansible, to build and maintain robust data pipelines and infrastructure.
  • Utilized Agile project management methodologies, with a focus on clear communication and task management using Linear, to ensure on-time and within-budget project delivery.
  • In close contact with stock optimization and configurability teams, I assisted in delivering the right data to enable sufficient functionality for client demos.