Featured
I spent my summer in Addis Ababa, Ethiopia 🇪🇹
I was part of the amazing team that ran AddisCoder 2024. Read more about my journey of a month in Ethiopia, where we taught the country's most talented high schoolers write their first lines of code in Week 1 and solve complex dynamic programming problems in Week 4.
Research
2024
Size Does Not Matter - A Data-Centric Approach to Fine-Tuning Phi-3
Sillystill - Recreating the look of Cinestill-800T using Deep Learning
2023
Advancing Homepage2Vec with LLM-Generated Datasets for Multilingual Website Classification
Unlocking the Palate - A Data-Centric Approach to Understanding Beer Descriptors
Benchmarking SOT Feature Transform for Biomedical Few-Shot Learning Tasks
Navigating Indoors with Computer Vision - Deep Learning Approaches for Room-Level Indoor Localisation
2022
Weighted Bipartite Graph Projection Methods for a GitHub Recommender System
Projects & Initiatives
TL;DR 📮
I enjoy reading academic papers and well-written technical blog posts. I capture the key ideas and takeaways in five concise bullet points. Sometimes, five is not enough - in that case, there is a Too Short; Want More (TSWM) button to get a more detailed summary.
Distributed Inference and Fine-tuning of Large Language Models Over The Internet
HSE•Borzunovet al.
(2023)
SWARM Parallelism - Training Large Models Can Be Surprisingly Communication-Efficient
HSE•Ryabininet al.
(2023)
Decentralised Training of Foundation Models in Heterogeneous Environments
ETH•Yuanet al.
(2022)
Reach me
© Mika Senghaas 2024