AI Dev & Researcher

Think Deep. Build Smart.

Summary

Machine Learning Engineer with 20+ years of experience in deep learning, LLMs, and AI systems. PhD in AI, founder of ML consultancy, and proven track record delivering production AI solutions for Fortune 500 companies. Expertise spans from cutting-edge research to scalable deployment, with strong focus on generative AI and model optimization.

Key Achievements: Reduced ML false positives by 40 percentage points • Trained 20+ production models • Led teams of up to 7 engineers • Won international AI competitions

📄 Download PDF Version

Education


October 2020 – September 2025	Degree in Mathematics. UNED
October 2011 – November 2012	MBA Part-Time at EOI (Escuela de Organización Industrial) Management School. Funded by EOI (see awards and prizes below). International stay at UCF for a week.
October 2008 – October 2011	Ph.D. in Computer Science and Artificial Intelligence. University Carlos III of Madrid. Research Focus: Imitation learning in video games and the development of believable NPCs Published papers in international conferences (CIG 2009, CIG 2010, UKCI 2010) Won the botprize in 2010 (Turing Test for videogames) with a Conscious Machine cognitive architecture
October 2007 – September 2008	Master in Computer Science and Information Technology. Specialization: Artificial Intelligence. University Carlos III of Madrid. Funded by the university.
October 2001 – November 2006	Computer Science Engineering (5 years career). Specialization: Artificial Intelligence. Graduated with first-class honors and the best bachelor's thesis. University Carlos III of Madrid.

Experience


February 2024 – May 2025	Machine Learning Engineer at Octane Security (Remote, California, US) LLMs for Smart Contract Vulnerability Detection: As the first ML hire at a blockchain security startup, led development of a multi-stage LLM-based system with over 20 models covering detection, explanation, patch generation, and verification. Reduced false positives from 50% to less than 10%. Helped scale the company from 3 to 10 people.
December 2021 – June 2022	Machine Learning Engineer at Realm Living (Remote, Argentina) 3D Point Cloud Instance Segmentation: Developed prototype for large-scale indoor environment analysis, classifying points into semantic objects and grouping them into instances.
January 2020 – now	Freelance Deep Learning Engineer/Researcher at Toptal / Upwork / 10x / Others: Edge AI & Model Optimization: Developed lightweight deep learning models achieving 95%+ accuracy for real-time applications. Optimized face embedding and adversarial detection models, reducing inference time by 85% and model size by 80% for mobile deployment Generative AI for Personalized Content: Built production-ready image generation pipeline using LoRA-based finetuning of Stable Diffusion models, achieving sub-60 second personalization. Led 6-month GAN training project for facial feature editing, delivering photorealistic results 3D World Generation with RL: 5-month R&D project training reinforcement learning agents to generate 3D environments using Houdini. Developed novel agent architecture operating GUI interfaces directly from natural language prompts Privacy-Preserving Face Synthesis: Created GAN-based system for synthetic face generation maintaining demographic characteristics while ensuring privacy compliance for identity protection applications Automated Content Assessment: Built deep learning model for image quality ranking with 80%+ correlation to human evaluators, processing 3M+ images daily for content platforms
April 2020 – September 2020	Machine Learning Engineer at NeuralX Limited (Remote) Stock price prediction models using wavelets, transformers, sequence to sequence models.
May 2018 – January 2020	Founder at Serendeepia In Serendeepia Research we were a small group of experts in Machine Learning and Deep Learning who built custom solutions based on state of the art research in Artificial Intelligence. I was CEO and CTO of the company. I created a remote-first culture with a well defined methodology to manage the projects: scrum and kanban boards. I picked our tools based on the best industry standards: Github with gitflow workflow for code, Kubernetes as cluster with Jenkins for continuous integration and Polyaxon as the platform to run experiments. I grew the company from 3 to 7 people, I managed successful projects for Fortune 500 companies and I lead the development of B2C products with the latest advancements in Deep Learning.
January 2016 – August 2018	Remote Machine Learning Engineer at Chute (San Francisco, USA). Solely responsible for processing social media data (Instagram, Twitter) in real-time for valuable insights. Chute was acquired in August 2018. Relevant projects: Multilabel image classification with TensorFlow and TensorFlow Serving (previously image classification with Caffe) in Python Aesthetic model to score the beautifulness of the images in TensorFlow Model in TensorFlow to create a perceptual hash of images using local and semantic information of images Library in Python to train, test and deploy TensorFlow and Caffe models Spam filter and user topics detection with Word2Vec and kmeans in Go Webapp in Ruby on Rails to organize the data for the image classification
June 2017 – April 2018	Remote contractor as Machine Learning Researcher at Good AI Lab (San Jose, CA, USA). I was in charge to create Deep Learning research projects that run in ClusterOne for promoting and testing the platform. For example: Personalized Medicine: Redefining Cancer Treatment with deep learning
March 2013 – September 2017	Android lead developer in startups in London (Quipper, Shopcade) and as a remote contractor (Appgree). Created AppHunt as a side project.
December 2010 – February 2013	Entrepreneur period. Startups and side projects with colleagues: Comaware: new technology based on Machine Consciousness to control virtual characters in video games and autonomous robots. Howmaths: Artificial Intelligence applied to math problems. An A* algorithms applied to solve math equation systems. A video game engine for Android developed from scratch as a side project to use Artificial Intelligence to control the non-player characters and procedural content generation.
September 2009 – January 2010	Visiting research at Center for Computer Games Research IT University of Copenhagen (Denmark), group of Artificial Intelligence in video games. The project realized was to control characters in video games throw an eye tracking system and Artificial Intelligence.
October 2007 – September 2011	Research staff in training at University Carlos III of Madrid. I supervised several bachelor's thesis. Rated by the students as an exceptional teacher (2008).
December 2006 – August 2007	Software developer at GTD Ingeniería de Sistemas y Software and Secosol soluciones y domótica.

Technical Skills

AI/ML Frameworks & Libraries Expert

PyTorch, TensorFlow, Transformers (Hugging Face), Stable Diffusion, LangChain, Sentence Transformers, scikit-learn, XGBoost, ONNX, TensorFlow Serving

Generative AI & LLMs Expert

OpenAI API, Claude API, Llama, Mistral, LoRA fine-tuning, RAG systems, Vector Databases (Pinecone, Weaviate, ChromaDB), Prompt Engineering

Programming Languages Advanced

Python, C++, Go, Kotlin, Java, Ruby, TypeScript, JavaScript, Bash

Infrastructure & Cloud Advanced

AWS (EC2, S3, Lambda, SageMaker), GCP (Compute Engine, Cloud Run, Vertex AI), Docker, Kubernetes, Jenkins, GitHub Actions, Terraform

Data & Tools Advanced

PostgreSQL, MongoDB, Redis, Git, Jupyter, Weights & Biases, MLflow, Polyaxon, FastAPI, Flask, Ruby on Rails

Awards and prizes


2011	Winner team of the Human-Like Bot contest. Winner project (Comaware) in the national contest for university entrepreneurs organized by EOI. Awarded with an MBA at EOI and a study trip to Silicon Valley. Winner team of the 5th contests of ideas for creating scientific and technological based companies of the University Carlos III of Madrid with the project Comaware. Six months of incubation in the business incubator of University Carlos III of Madrid.
2010	Winner team of the 2K BotPrize 2010 international contest. The 2K BotPrize is an adaptation of the Turing test to video games. Monetary prize of a $2,000 and a trip to 2K studios in Canberra, Australia. Winner of the Car Set-up Optimization competition (using TORCS as simulator). $500 prize.
2008	Rated by the students as an exceptional teacher, congratulated by the University Carlos III of Madrid
2006	Graduated with first-class honours and best bachelor's thesis
2005	Scholarship for outstanding academic achievement in the course 2005-2006.
1999	Finalist in the fourth edition of the national Spring Mathematics Contest held at the University Complutense of Madrid.
1998	Finalist in the third edition of the national Spring Mathematics Contest held at the University Complutense of Madrid.

Media Appearances


October 2010	Killer Bots Are Getting Human. John Bohannon. Science. Vol 330. no. 6000. pp. 30-31. Science
October 2010	The Awakening of Conscious Bots: Inside the Mind of the 2K BotPrize 2010 Winner, aigamedev.com.
September 2010	Unreal Tournament 2010: Narrowing the Gap between Human and Bot. Surfdaddy Orca. h+ magazine.
September 2010	Un equipo español queda en primer lugar del 2K BotPrize 2010 (A Spanish team win the 2K BotPrize 2010). Lektu / Microsiervos. lainformacion.com

Talks


March 2011	Guest speaker in the VI Days of Imaginática 2011. University of Sevilla. Conciencia Artificial en Videojuegos: diseñando máquinas más humanas. (Machine consciousness in video games: designing more human-like machines).

Want to get in touch?

I'm open to collaborations, project proposals, or just a chat about AI.

Contact me