ActiveUltraFeedback: Efficient Preference Data Generation using Active Learning
Development of ActiveUltraFeedback, a RLHF pipeline for Preference Data Generation utilizing Active Learning 
Posts by Collection
portfolio
Bachelor Thesis: Exploring the Hidden Structures of Attention Layers in Transformer Models through the Lens of Gaussian Distributions (July 01, 2024)
Analyzing the mathematics of attention layers through random matrix theory and a finite-dimensional Gaussian approximation 
ETH Large-Scale AI Engineering 2026: Mamba, DeltaNet and Torch DDP
Custom implementation of the DeltaNet and Mamba SSM, comparing performance/throughput against a Transformer baseline, and testing distributed data parallel (DDP) schemes 
ETH Computational Intelligence Lab 2025: Uncertainty-Aware Ensemble for Monocular Depth Estimation
Fine-tuning a mixture-of-experts meta-model for monocular depth estimation using epistemic uncertainty estimates 
publications
ActiveUltraFeedback: Efficient Preference Data Generation using Active Learning
Published in International Conference on Machine Learning (ICML), 2026
* Equal contribution
ACTIVEULTRAFEEDBACK introduces a modular active learning pipeline that uses uncertainty-aware reward estimates to select informative response pairs, reducing the amount of preference data needed for strong downstream performance.
talks
Talk 1 on Relevant Topic in Your Field
Published:
This is a description of your talk, which is a markdown file that can be all markdown-ified like any other post. Yay markdown!
Conference Proceeding talk 3 on Relevant Topic in Your Field
Published:
This is a description of your conference proceedings talk, note the different field in type. You can put anything in this field.
teaching
Teaching experience 1
Undergraduate course, University 1, Department, 2014
This is a description of a teaching experience. You can use markdown like any other post.
Teaching experience 2
Workshop, University 1, Department, 2015
This is a description of a teaching experience. You can use markdown like any other post.