Research Scientist at KAUST

My research interests span Computer Vision, Machine Learning, AI Safety, and AI for Science. I focus on developing safe, accurate, and efficient vision systems and machine-learning models that can reliably perceive, understand, and act on information, while ensuring robustness, protecting privacy, and aligning with societal values.

I received my B.Sc., M.Sc., and Ph.D. degrees in Computer Science from Universidad Industrial de Santander, Colombia, in 2015, 2018, and 2022, respectively. I was a research intern at the Stanford Vision and Learning Lab (SVL), supervised by Prof. Juan Carlos Niebles. I am currently a Research Scientist at KAUST in the Image and Video Understanding Lab (IVUL), led by Prof. Bernard Ghanem, where I lead the AI Safety research theme.

Resume Contact me

📣 Open positions. I'm looking for Ph.D. students and postdocs to join IVUL at KAUST. You will work with me and Prof. Bernard Ghanem on AI safety, with a focus on the reliability, robustness, trustworthiness, and security of AI models and agentic systems. I am also looking for Ph.D. students interested in research internships at KAUST. To apply, please fill out this Google form; I read every application and will get in touch if there is a good fit. If you have a research idea that connects with my interests and would like to collaborate, feel free to fill out the form or email me at carlos DOT hinojosa AT kaust DOT edu DOT sa.

Highlights

| Oct. 2026 | I have been promoted to Research Scientist at KAUST! I will continue leading the AI Safety research theme in IVUL. I’m also looking for PhD students and postdocs to work with me on AI safety. If you are interested, please fill out this Google form.
| Oct. 2026 | Two U.S. patents from my research on privacy-preserving computer vision and computational imaging, a collaboration between Stanford University and Universidad Industrial de Santander (UIS), were issued this year:
- 📃 US 12,651,379 B2: Systems and Methods for Privacy-Preserving Optics. Inventors: Juan Carlos Niebles, Carlos Hinojosa, and Henry Arguello.
- 📃 US 12,738,096 B2: Systems and Methods for Recognizing Human Actions from Privacy-Preserving Optics. Inventors: Juan Carlos Niebles, Carlos Hinojosa, Henry Arguello, Miguel Marquez, Ehsan Adeli, and Fei-Fei Li.
| Jul. 2026 | Our paper “Structured-Noise Masked Modeling for Video, Audio and Beyond” will appear as an oral presentation at ECCV 2026 in Malmö, Sweden! Check out the project page.
| Jul. 2026 | Our paper “Multimodal Safety Evaluation in Generative Agent Social Simulations” was published at ACL 2026!
| Jun. 2026 | I received a Gemini Academic Program Award from Google! The award provides USD 20,000 in Google Cloud credits for Gemini to support my AI safety research at KAUST.
| Jun. 2026 | Our paper “CURE: Curriculum-Guided Multi-Task Training for Reliable Anatomy-Grounded Report Generation” was presented as an oral presentation at CVPR 2026!
| May. 2026 | I started an academic visit at the University of Oxford, working closely with Adel Bibi and Philip Torr, FREng, FRS on AI safety research.
| Apr. 2026 | Excited to share that our Competitive Research Grant (CRG) proposal on Safe Operating System Agents has been awarded! Over the next three years, this KAUST-Oxford collaboration (with Prof. Bernard Ghanem and Prof. Mohamed Elhoseiny at KAUST, and Dr. Adel Bibi and Prof. Philip Torr at Oxford) will focus on building safe and reliable AI agents that operate across real-world digital environments.
| Feb. 2026 | I presented our paper “ColorMAE: Exploring Data-Independent Masking Strategies in Masked AutoEncoders” at the Research Symposium on AI & Its Impact, part of the India-AI Impact Summit 2026 in New Delhi, India.
| Feb. 2026 | I served as Session Chair and Scientific Committee member for the KAUST Rising Stars in AI Symposium 2026.
| Jan. 2026 | I am an invited speaker at The Middle East and North Africa Machine Learning (MenaML) Winter School 2026. Talk title: “The Rise of Multi-Agent Systems: Capabilities, Opportunities, and Safety Considerations”.
| Jul. 2025 | Our Paper UnMix-NeRF has been accepted in ICCV 2025!
| Jun. 2025 | I’m excited to share that I’m co-organizing the 1st Workshop on Safe and Trustworthy Multimodal AI Systems (SaFeMM-AI) at ICCV2025!
| May. 2025 | I am an Area Chair for BMVC 2025.
| Sep. 2024 | I am an invited speaker at the 2nd Workshop on Privacy For Vision & Imaging at ECCV 2024.
| Jul. 2024 | Our ColorMAE paper has been accepted in ECCV 2024, and will also be presented at the Self Supervised Learning: What is Next workshop! See you in Milan.
| Aug. 2023 | I am editor of the EURASIP Journal on Image and Video Processing. Check the published Collection articles here
| 2021 - 2023 | I co-organized the LatinX in Computer Vision (LXCV) workshop at ICCV 2021, CVPR 2022, CVPR 2023, and ICCV 2023.

| Oct. 2022 | Our paper “Optics Lens Design for Privacy-preserving Scene Captioning” won the Best Paper Award from the IEEE International Conference on Image Processing (ICIP2022)! See here.
| Apr. 2022 | I successfully defended my Ph.D. and got the laureate distinction!

Selected Publications

APEX-Harm: Evaluating Agent Safety in Shared Professional Workspaces

arXiv 2026

Ziyi Yang, Adel Bibi, Guohao Li, Bernard Ghanem, Carlos Hinojosa

PDF

Structured-Noise Masked Modeling for Video, Audio and Beyond

ECCV 2026 (Oral Presentation)

Aritra Bhowmik, Carlos Hinojosa, Fida Mohammad Thoker, Bernard Ghanem, Cees G. M. Snoek

PDF Code Project

Multimodal Safety Evaluation in Generative Agent Social Simulations

ACL 2026 (Main Conference)

Alhim Adonai Vera Gonzalez, Carlos Hinojosa, Karen Sanchez, Haidar Bin Hamid, Donghoon Kim, Bernard Ghanem

PDF Code Project

CURE: Curriculum-Guided Multi-Task Training for Reliable Anatomy-Grounded Report Generation

CVPR 2026 (Oral Presentation)

Pablo Messina, Andrés Villa, Juan León Alcázar, Karen Sanchez, Carlos Hinojosa, Denis Parra, Álvaro Soto, Bernard Ghanem

PDF Code Project

ColorMAE: Exploring data-independent masking strategies in Masked AutoEncoders

ECCV 2024

Carlos Hinojosa, Shuming Liu, Bernard Ghanem

PDF Supplementary Project Poster DOI

PrivHAR: Recognizing Human Actions from Privacy-preserving Lens

ECCV 2022 | European Computer Vision Association (ECVA)

Carlos Hinojosa, Miguel Marquez, Henry Arguello, Ehsan Adeli, Fei-Fei Li, Juan Carlos Niebles

PDF Supplementary Project Poster Video

Optics lens design for privacy-preserving scene captioning

ICIP 2022 | IEEE Conference

Paula Arguello, Jhon Lopez, Carlos Hinojosa, Henry Arguello

PDF Project Poster Video

Learning Privacy-preserving Optics for Human Pose Estimation.

ICCV 2021 | IEEE Conference

Carlos Hinojosa, Juan Carlos Niebles, Henry Arguello

PDF Supplementary Code Project Poster Slides Video

See all publications

Talks

Invited talks and keynotes

Carlos Hinojosa speaking at the MenaML Winter School 2026

2 photos

Jan 2026 · MenaML Winter School 2026 · KAUST, Saudi Arabia

The Rise of Multi-Agent Systems: Capabilities, Opportunities, and Safety Considerations

Carlos Hinojosa giving a keynote at the ECCV 2024 Privacy for Vision and Imaging Workshop

3 photos

Sep 2024 · Privacy for Vision & Imaging Workshop at ECCV · Milan, Italy

Privacy-preserving in Computer Vision through
Optics Learning

Feb 2026
Exploring Data-Independent Masking Strategies in Masked AutoEncoders
Research Symposium on AI & Its Impact, India–AI Impact Summit 2026 · New Delhi, India
Jan 2026
The Rise of Multi-Agent Systems: Capabilities, Opportunities, and Safety Considerations
Middle East and North Africa Machine Learning (MenaML) Winter School 2026 · KAUST, Saudi Arabia
Nov 2025
Privacy-preserving in Computer Vision through Optics Learning
Invited speaker, Simposio Internacional de Física (SIF) 2025 · Universidad de Pamplona, Colombia
Sep 2024
Privacy-preserving in Computer Vision through Optics Learning
Keynote, 2nd Privacy for Vision & Imaging Workshop, ECCV 2024 · Milan, Italy
2023
Privacy-preserving in Computer Vision through Optics Learning
Rising Stars in AI Symposium 2023 · KAUST, Saudi Arabia
2023
Privacy-preserving in Computer Vision through Optics Learning
Computational Imaging Group, KAUST Visual Computing Center · KAUST, Saudi Arabia
2022
PrivHAR: Recognizing Human Actions From Privacy-preserving Lens
Stanford Vision and Learning Lab (SVL) · Stanford University, USA
2021
Systems and Methods for Privacy-preserving Computer Vision
Stanford Vision and Learning Lab (SVL) · Stanford University, USA