Yuliy Tsank Email & Phone Number

San Francisco Bay Area

• Developed ML models and workflows for a broad range of NASA, FAA, and DoD automation use cases with a wide range of data types (images, text, audio, tabular) ◦ Object detection on airfield surfaces to avoid accidents and engine damage ◦ Biometric iris and voice identification for active environments in the field ◦ NASA space observatory time series processing for efficient data collection ◦ Audio data processing for efficient army radio testing ◦ Free-form text processing for condition-based aircraft maintenance standardization ◦ Prediction of airport runway capacity for efficient air traffic operations ◦ Obstacle avoidance in simulated autonomous vehicles - reinforcement learning• Wrote technical proposals for a wide range of ML applications for government entities via the SBIR (Small Business Innovation Research) program ◦ Quickly researched state-of-the-art ML methods via literature reviews ◦ Expanded company capabilities into new areas of AI/ML

San Francisco Bay Area

• Completed an intensive postdoctoral industry-training program • Built web app to automate glaucoma detection from 3D medical images (Ocular Coherence Tomography) with ~90% accuracy (.9 AUC-ROC)• Built and trained custom neural network model based on a recent medical imaging paper to output glaucoma probability and guide manual diagnosis in difficult cases with attention-heatmaps• Extracted 3D class activation maps from trained network to find discriminative regions in image volumes • Used model as proof of concept for automated glaucoma detection with up to 15x less costly equipment and significantly less human input for potential use in the developing world

Santa Barbara, California Area

𝐂𝐨𝐦𝐩𝐮𝐭𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐕𝐢𝐬𝐢𝐨𝐧 𝐒𝐜𝐢𝐞𝐧𝐜𝐞 𝐋𝐚𝐛• Deep Learning (Neural Network) Models: implemented and optimized VGG, Resnet, and various small CNN models along with tuning hyperparameters, trained models from beginning for visual classification tasks, used models to study human neural representation of faces• Statistical Models: implemented various Bayesian Ideal Observer models, developed new optimal eye-movements model for videos of faces, developed new optimal eye-movements model for simulated visual impairment, fit models to human experimental data• Experimental Design with Psychophysics and Eye-tracking: implemented visual perception studies, developed a simulation of visual impairment in a gaze-contingent paradigm• Programming and Data Analysis in Matlab and Python: analyzed behavioral data and images with computer vision and numerical toolboxes, developed computational models using statistical toolboxes, developed experiments with Psychtoolbox and interfaced with an eye-tracker• Gave talks and presented posters at Vision Sciences Society conferences • Taught/Assisted courses in: Perception Lab and Lecture, Intro Psychology, Statistics

Greater San Diego Area

𝐂𝐨𝐦𝐩𝐮𝐭𝐚𝐭𝐢𝐨𝐧𝐚𝐥/𝐂𝐨𝐠𝐧𝐢𝐭𝐢𝐯𝐞 𝐍𝐞𝐮𝐫𝐨𝐬𝐜𝐢𝐞𝐧𝐜𝐞 𝐋𝐚𝐛• Designed visual perception experiment and Bayesian model to study how people integrate information over time during perceptual decision-making

Doctor Of Philosophy - Phd, Psychological And Brain Sciences

Bachelor'S Degree, Cell/Cellular And Molecular Biology