At Shipd (Eris), I created original ML/AI challenges with problem statements, datasets, scoring metrics, and baselines, and submitted solutions across NLP, computer vision, graphs, and scientific ML. I ranked Top 50 overall, with a 1st-place result and multiple top-3 finishes.
At AfterQuery, I designed and tested evaluation tasks for coding agents, including Harbor/Terminal-Bench sys-admin tasks for Project Pluto and real-bug tasks for Project Silver. Earlier, I developed backend systems, APIs, and database structures at ZalaTech and University of Gondar (UOG).

