Career & education
Research, engineering, and teaching across foundation models and multilingual NLP.
Download CV PDFExperience & Education
Senior LLM Researcher
- Training large language models, from data acquisition and pipeline development to pre-training, fine-tuning, reinforcement learning, and optimization at scale.
LLM Researcher
- Developing compressed LLMs that make AI systems faster, cheaper, and more energy-efficient, including Hypernova-60B, a 60B-parameter model optimized for agentic workflows and tool use.
- Developed a steering-vector method for editing LLM knowledge and behavior, including adding or removing censorship around sensitive topics while preserving model performance.
- Optimized internal inference and fine-tuning codebases, achieving 3× faster logit precomputation and 5× faster KL-divergence training in internal pipelines, relative to their prior implementations.
- Built large-scale synthetic data generation pipelines for LLM training and evaluation.
Machine Learning Researcher
- Contributed to the development of FLUX.1 Krea, a diffusion-based image model trained in collaboration with Black Forest Labs.
- Developed a personalized image recommendation engine using user interaction data and deep learning techniques.
- Designed and trained SigLIP-style embedding models for artistic style-based image retrieval.
- Improved fine-tuning codebases for video (Wan 2.1) and image (Flux) models, focusing on training performance and distributed training techniques such as FSDP2.
- Implemented and deployed generative services for 3D models and textures, powering features used by thousands of users on the Krea platform.
Visiting PhD Student
- Advisor: Dan Roth
- Developed multilingual named entity recognition using Wikipedia and Wikidata, and worked on T-Projection for cross-lingual annotation projection.
PhD in Natural Language Processing
- Title: Acquisition and exploitation of Cross-Lingual knowledge
- Advisors: German Rigau, Rodrigo Agerri
Applied Scientist Intern
Master's in Natural Language Processing
Bachelor in Computer Science
Grants and Awards
Exponential Fellow
The Exponential Fellowship: A fellowship program that aims to find brightest builders in Spain to send on a life-changing experience to the US.
https://www.goexponential.org/directory/fellows/iker-garcia-ferrero
Winner of the ODESIA Challenge @ SEPLN 2024
Winner of the ODESIA Challenge. The aim of the competition is to promote the development and evaluation of language models in Spanish.
Best Project Award @ SomosNLP 2024 Hackathon
Best Project Award in the SomosNLP 2024 Hackathon for Clickbait Article Summarization in Spanish.
Egonlabur: Grants for stays in centers other than the one in which the PhD is carried out
Egonlabur grants for stays in centers other than the one in which the PhD is carried out. 3.475€
Prize for the most relevant research for the development of the Basque Country (IkerGazte 2021)
The paper "Twitterreko Euskal Komunitatearen Eduki Azterketa Pandemia Garaian" was awarded the prize for the most relevant research for the development of the Basque Country at IkerGazte 2021.
Basque Government PhD grant
4 year PhD grant.
Grants for university master's degrees at the UPV/EHU
Amount: 3000€.