DIEGO BONILLA SALVADOR / PORTFOLIO
Welcometo mywork.
Deep learning researcher from Spain. I build and study computer vision, generative models, and tools that make AI easier to understand.
DIEGO BONILLA SALVADORPROJECTS
ARTICLES
VIDEOS
ARTICLES
VIDEOS
DIEGO BONILLA SALVADOR / PORTFOLIOSCROLL DOWN ↓
Browse by section.
SEARCH THE SITE ↗
Selected projects.
ALL PROJECTS ↗A few projects from the archive.
ROBOTICS / COMPUTER VISIONQuadruped Robot Object Detection Person FollowingA quadruped-robot demo that uses YOLOv11 detections to identify a person and drive a follow behavior in real time. The project combines onboard perception with mobile-robot control.VIEW PROJECT COMPUTER VISION / MULTIMODAL AIPixLore: A Dataset-Driven Approach to Rich Image CaptioningPixLore investigates richer image captions by fine-tuning BLIP-2 with LoRA on a curated and augmented dataset. The project asks whether smaller specialist captioners can contribute complementary detail to image understanding.VIEW PROJECT REINFORCEMENT LEARNINGReinforcement Learning: Proximal Policy OptimizationA from-scratch Proximal Policy Optimization implementation trained on Flappy Bird, with Snake explored as a next game environment. The project makes the policy-learning loop tangible through game play and agent behavior.VIEW PROJECT COMPUTER VISIONFractalTokenizerAn adaptive image tokenizer that combines corpus-level variable-byte-pair encoding with fractal quadtree partitioning. It creates variable-size patches intended to balance spatial detail with the information each patch carries.VIEW PROJECT COMPUTER VISION / MULTIMODAL AISceneGraph-LLMGenerates scene graphs from images by combining LLM-created synthetic supervision with a vision-language model for object relationships. The output represents detected entities and their relationships as structured graph data.VIEW PROJECT MULTIMODAL AIPruneForge: Zero-Training LLM Pruning for AgentsPruneForge searches for a task-specific, smaller language model by removing MLP channels, attention heads and layers while checking output similarity against a small prompt set. The pruning loop uses no fine-tuning and saves an inspectable pruning recipe.VIEW PROJECT
A book I coauthored.
AI Revolution: Exploring Advanced Concepts in Deep Learning
A coauthored textbook that pairs deep learning theory with practical laboratory exercises, covering neural networks, transformers and LLMs, generative models, reinforcement learning, optimization, and deployment.
Coauthored with Emilio Soria Olivas, Pablo Rodríguez Belenguer, Emma Amorós Belda, Pablo Hernández Cámara and Manuel Sánchez-Montañés.
VIEW BOOK ↗Articles and videos.
A selection of things I've written and recorded.
ARTICLES
01Tokenization in NLP: An Academic GuideAn introduction to how language models turn raw text into tokens and numerical representations, including word and subword units and Word2Vec-style embeddings.02A Guide to Transformer ArchitecturesA practical guide to transformer building blocks, self-attention, and architecture variants across language and computer vision, with attention to how the pieces fit together.BROWSE ARTICLES ↗VIDEOS
01How to Make Your LLM ReasonA Spanish-language video in Diego’s series on deep learning methods for language models, focused on techniques for eliciting and training reasoning behavior.02Real-Time RGB-D SLAMA real-time demonstration of camera tracking and 3D mapping from Intel RealSense D415 depth data using OpenCV and Open3D.BROWSE VIDEOS ↗Music, too.
Two music projects from the old portfolio.