- Face recognition in real-world scenarios- Person re-identification - Designed and implemented multi-object tracking systems for real time

- Define metrics for community immunity.- Predict pandemic outbreaks using sewage epidemiological methods, as an efficient alternative to individual testing and, build a dashboard for visualization.

- Represent knowledge structure as complex networks.- Build Python packages for network analysis (Ricci curvature and flow) and novel local centrality measure. Applications: next word prediction, contextual representation of objects

First data scientist in MAGNA (HUJI spinoff), an ML diagnostics pre-seed startup utilizing speech-based signals (NLP & voice) for diagnosing psychopathology conditions.- Algorithm development: Feature extraction and verification of pathological thought patterns through voice and NLP for diagnosis of ADHD, PTSD and schizophrenia.- Developed a proprietary Python package for time-series analysis and feature extraction of networks.- POC: Developed Neural networks and classical ML models, achieving F1 > 0.9 for ADHD and schizophrenia classification (baseline F: ~0.5). Currently in development under Israel Innovation Authority grant (הרשות לחדשנות)- Lead data pipeline creation: app development, transcription (praat, sonix), AWS (S3 & DynamoDB), preprocessing.