Vukasin Bozic

Stefano Franscini Platz 5, ETH Zürich · 8093, Zürich, Switzerland · vukasin.bozic@ethz.ch
PhD Student @ ETH Zürich | x-Research Intern @ Disney Research | ETH Zürich Master Graduate

I am a doctoral researcher at the Photogrammetry and Remote Sensing Lab, ETH Zürich, specializing in 3D Computer Vision, 3D Reconstruction, and Diffusion Models. My journey spans academia, teaching, cutting-edge research, and industry experience, including roles in both established organizations and innovative startups. I am deeply passionate about advancing Deep Learning and Artificial Intelligence while exploring their practical applications in the real world.

Publications

  • Versatile Vision Foundation Model for Image and Video Colorization
    V Bozic, A Djelouah, Y Zhang, R Timofte, M Gross, C Schroers
    ACM SIGGRAPH, 2024
    Link to paper

  • Rethinking Attention: Exploring Shallow Feed-Forward Neural Networks as an Alternative to Attention Layers in Transformers
    V Bozic, D Dordevic, J Thommes, D Coppola, SP Singh
    Proceedings of the AAAI Conference on Artificial Intelligence, 2024
    Link to paper

Experience

Doctoral Researcher

ETH Zürich

Doctoral studies at Photogrammetry and Remote Sensing group at ETHZ, projects revolving around Computer Vision, 3D Vision and Deep Learning. Currently working on a projects in 3D Mesh Simplification and Optimization, and Diffusion Models.

November 2024 - present

Research Intern

Disney Research Zürich

Deployed and improved on state-of-the-art Monocular Depth Estimation models for reconstruction of 3D scenes in Stereo Conversion procedures, enabling simple 2D scenes and movies for 3D technology and VR experience.

February 2024 - September 2024

Machine Learning Engineer

The Fire Company

Automated the complex and demanding process of brand and product curation and refinement in the context of e-commerce business. Developed the cloud functions in Google Cloud Platform which interacted with Google Firestore and Storage. Designed complex pipelines for data processing and comprehension using Generative AI by exploiting the power of Firebase Genkit plugins and prompt engineering. Developed and maintained the Google Genkit plugins for OpenAI, Groq, AnthropicAI, MistralAI...

February 2024 - July 2024

Master Thesis Intern

Disney Research Zürich

Master thesis project done during the stay at Disney Research was published at the prestigious ACM SIGGRsAPH24 conference. The work exploits huge pre-trained text-to-image Diffusion Models for the task of image and video colorization. It achieves high-quality, State-of-the-Art results, along with highly controllable, efficient, and customizable colorization.

February 2024 - September 2024

Graduate Teaching Assistant

ETH Zürich

Involved in courses Probabilistic Artificial Intelligence and Introduction to Machine Learning. Worked on design of ML and AI student projects, and development of automatic submission and evaluation interface. Implemented creative and challenging coding projects, where students engaged in implementation of newest methods in field of Probabilistic AI. Improved cross-platform automatic evaluation interface using Docker, enhancing user experience and streamlining results submission process

September 2022 - January 2024

Signal Processing and Control Intern

Zanus Technology Belgrade

Actively contributed to the research and development of specialized PTZ camera head for video production automation. Focused on the design of automated tracking algorithms for precise control of PTZ and PTU camera heads, ensuring highquality and stable camera operation, in scope of Bachelor thesis. Implemented gyro-stabilization control algorithm using IMU device, enhancing camera head stability across diverse scenarios. Projects spanned from low-level microcontroller programming to high-level deep learning-based tracking algorithms.

February 2020 - July 2021

Education

ETH Zürich

Doctorate
Photogrammetry and Remote Sensing, Doctorate Studies
November 2024 - present

ETH Zürich

Master of Science
Information Technology and Electrical Engineering Master

GPA: 5.6/6.0

September 2021 - February 2024

University of Belgrade, School of Electrical Engineering

Bachelor of Science
Signal Processing and Control

GPA: 9.8/10.0

September 2017 - July 2021

Skills

Python C++ PyTorch, Scikit-learn, Open3D Git, Linux, Docker Machine Learning & AI Deep Learning Computer Vision 3D Computer Vision & Reconstruction Diffusion Models SfM, SLAM Cloud Computing, GCP JavaScript, TypeScript

Accomplishments