Liel Amar
📅 Present
I've been building my own and others' ideas since I was 13, and I'm passionate about software design & architecture. My expertise lies in bringing product visions to life, mostly through mobile and web apps.
Currently, I am pursuing my Master's degree in Computer Science, under the Supervision of Prof. Yedid Hoshen. My main focus is Weight Space Learning.
I am 24 years old and I currently live in Israel. Offline, you might find me shooting videos, going to the cinema, watching my favorite teams play or traveling the world!
🎓 Education
🖥️ Experience
I designed and implemented solutions for data storage, machine learning training & tuning, and client interaction. I introduced the company's first CI/CD pipelines, automating unit testing and linting. I also conducted code reviews to fellow developers.
• The company ceased operations in September 2024 due to financial difficulties.
I developed a real-time audio streaming service which connected a public web application, staff moderation tools and video games, used to host virtual events.
• I left the company to pursue my Bachelor's.
🛠️ Projects
Double You
Lead Developer
A self-hosted context-aware personalized AI assistance mobile application using Retrieval Augmented Generation. This project won HUJI's 2025 Hackathon.
Defeating Frogger with AI
Lead Developer
Using Reinforcement Learning and Genetic Algorithms to create AI agents that beat the Frogger game.
Two Factor Authentication
Lead Developer
A two-factor authentication system for Bungeecord and Spigot servers. It has over 8,500 downloads on SpigotMC.
VolunTime
Lead Developer
An android application used for finding volunteer opportunities using LLMs. It was developed as part of HUJI's 2023 Hackathon.
LanguageFix
Lead Developer
A Spigot plugin that fixes RTL issues in Minecraft. It has over 1,500 downloads on SpigotMC.
Party Games
Lead Developer
A clone of Hypixel's Party Games. It contains 26 different Mini-games and supports full customization.
🏆 Awards
🏆 Publications
Arxiv Preprint
We propose the Model Atlas, an interactive visualization and analysis tool for exploring large-scale AI model repositories. To explore the hidden potential of model repositories, we chart a preliminary atlas representing the documented fraction of Hugging Face. It provides stunning visualizations of the model landscape and evolution. We demonstrate several applications of this atlas including predicting model attributes (e.g., accuracy), and analyzing trends in computer vision models. However, as the current atlas remains incomplete, we propose a method for charting undocumented regions.
