Pere Martra Manonelles

Pere Martra is a Research Engineer specializing in the optimization and ethical adaptation of Large Language Models. Author of Large Language Models Projects (Apress, 2024), he is currently developing novel pruning and adaptive inference techniques for his next book with Manning Publications.

His research focuses on structured pruning, model efficiency, and fairness-aware optimization. Creator of OptiPfair and WizardSData—open-source tools for model pruning, bias detection, and synthetic data generation respectively. His published research on GLU-based architectures demonstrates how optimization can bridge performance and environmental sustainability in AI.

As an active open-source contributor, Pere maintains GitHub repositories with over 2K stars, including a widely-adopted hands-on LLM course, and contributes to Hugging Face and Google Gemini Cookbooks. He holds a Master's Degree in AI Research.

Connect: For questions about the book or his optimization techniques, reach out via social media or the book's GitHub repository—Pere actively engages with the technical community.