David Font

AI/ML Engineer with hands-on experience building end-to-end LLM systems, RAG pipelines, and AI agents in production. I design, build, and ship — from scoping with stakeholders to deployment and iteration based on real user feedback. I use AI tools daily (Cursor is central to my workflow) and I'm always looking to go further with them. Based in Barcelona, available immediately.

ML Engineer at Multiverse Computing (2026-01 – 2026-03)

• Developed neural network compression and profiling pipelines for computer vision models (UNet, ConvNeXt, RegNetY) applied to satellite image processing.
• Worked in an international agile team using Cursor daily, GitLab CI/CD, and AWS EC2/S3.

ML Engineer at Novatronic S.A. (2024-09 – 2025-10)

• Identified the problem with QA and dev stakeholders, designed the full architecture, and shipped a production LLM + RAG system that automated test case generation from technical documentation.
• Built the complete pipeline: OCR ingestion, chunking strategy, sentence-transformer embeddings, ChromaDB vector store, FastAPI backend, Streamlit UI — deployed on AWS EC2 with Docker and GitLab CI/CD.
• Learned: how to design chunking and retrieval for real documents, how to evaluate RAG quality (RAGAS), and how to iterate a system in production based on user feedback. ~40 active users, ~30% reduction in manual effort.
• Integrated OpenAI enterprise API with structured prompt engineering, few-shot techniques, and safe-by-design patterns. Added human-in-the-loop feedback loop for continuous improvement.
• Designed and shipped an ML pipeline over ~1.5M monthly transactions for a regulated banking client.
• Learned: how to handle severe class imbalance (SMOTE), temporal splitting to avoid leakage, and threshold optimization for recall. XGBoost achieved AUC-ROC 0.93, ~25% reduction in false negatives.
• Built a fully auditable pipeline with explainability (SHAP values) to satisfy compliance requirements.

R&D Researcher at HP Inc. (2023-05 – 2024-01)

• Data analysis and scientific literature review (50+ publications) within a 4-person R&D team on a biodegradable materials project.

Master's in AI & Data Science in AI & Data Science – Universidad Europea de Madrid (UEM) (2023 – 2024)

Bachelor's Degree in Chemistry in Chemistry – Universitat de Barcelona (UB) (2014 – 2023)