▪ Agile developing of real time behavioral recognition system based on video understanding.▪ Research and Development of machine learning models and data driven heuristic solutions into production.▪ Experience with large scale cloud computing resources for training and models optimization.▪ Developing in-house tools for data analysis and model evaluation (analytical and visual).

▪ Member of the global AI core team which assist in introducing and leveragingmachine learning capabilities in Samsung/SNext digital health products and services.▪ Implementing and analyzing POC solutions to increase positive engagement in video interactions (e.g., gaze, activity and eye-movement, facial emotion).▪ Collaborating with Samsung research groups.

▪ Designing and Developing optimized machine learning solutions to be used by AI SOC low-cost processors with minimal memory and computing costs for OEMs and IoT solution providers.▪ Working with low resolution CMOS sensor data.▪ Developing classification decision tree-based ensemble models and deep learning CNN/regressor models for sensing human body presence of various nature (e.g., head, full body).▪ Developing embedded pipelines stages for motion and object detection based on the classification models.▪ Developing data science tools for annotations and pipeline evaluation.

▪ Developing full R&D cycle automated deep learning solutions for inspecting anomalies in a vehicle's (cars/trucks) undercarriage images.▪ Research and Development of supervised learning methods - classifiers, semantic segmentation and patch-based comparison models.

Developing background models for graph-based image segmentation of product images

Developing Knowit java-based web-application, an advanced automated marketing application for campaign planning. Knowit helps the marketeer to target groups of customers, make pre-counts, calculate business cases and prepare a full campaign brief.