welcome to my corner of the internet
CS student at BGU. I like asking hard questions,
then writing code until they have an answer.
Data science, deep learning, and algorithmic trading.
An autonomous Data Engineering and AI platform built to analyze personal Garmin watch data. The system solves LLM hallucination through a "Determinism before Semantics" architecture - mathematical ground truth is established before any AI touches the data.
A fully automated trading system that scans the NASDAQ-100 daily for mean-reversion and momentum signals, executes bracket orders through Interactive Brokers, and persists every trade, portfolio snapshot, and performance metric to a Postgres database.
A deep learning system that treats visual localization as a continuous regression problem rather than a discrete retrieval task - no GPS signal, just a photo and a prediction. Tested on BGU campus with 3,646 images captured in 360 degree rotations for viewpoint invariance.