Adrien Matea-Adan
I am a passionate AI/ML engineer and data scientist with extensive experience building machine learning products focused on production readiness and best practices.
About
I am a passionate AI/ML engineer and data scientist with extensive experience building machine learning products focused on production readiness and best practices. I specialize in developing AI-powered applications and platforms that enhance user experience and improve business outcomes. I enjoy working across the full stack of technologies including Python, AWS cloud services, and modern ML frameworks, combining my skills to deliver scalable and efficient AI solutions.
With a strong background in managing teams and projects, I excel at collaborating with stakeholders across technical and business domains to define objectives, optimize workflows, and deploy innovative AI products. I am continuously learning and adapting to the latest trends in AI, deep learning, and MLOps to drive successful and impactful implementations in real-world scenarios.
- 8 years' experience
- Joined August 2025
Languages
- EnglishAdvanced
- FrenchAdvanced
Industry experience
- Software & Internet
- Financial Services
- Media & Entertainment
- Professional Services
- Retail
Work experience· 4 roles
Managed 3 ML/AI engineers focusing on training, task definition, team progress, and code review. Developed an AI-powered chat and image generation platform using Python and AWS serverless technologies to enable context-aware conversations, voice input, and user management. Built an AI-powered social media management platform automating comment retrieval, contextual LLM responses, and content posting across platforms such as YouTube, Facebook, Instagram, and Google Business using FastAPI, Python, and PostgreSQL.
Managed 3 data scientists performing clustering for B2C and B2B clients to create targeted email marketing segments and identify high-value leads. Recalculated clusters monthly using classification models such as SVM and Random Forest. Developed mobile document classification for Digiposte iOS and Android apps, implemented with Mobile Net V2 and TensorFlow Lite to optimize production inference. Created distilled end-to-end visual document understanding models using Swin Encoder and BART Decoder, bypassing OCR for improved efficiency and accuracy.
Performed pre-launch testing of new cosmetic products globally, leading all project aspects from conception to client presentations using data analysis. Developed comprehensive data-driven presentations informing client decision-making.
Education
Certifications and awards
Hire Adrien Matea-Adan today
Post a job and invite Adrien, or contact them directly. No hidden commission.