Oct 2023 – Jul 2025 · 1 yr 9 mos · Remote

• Developed an EHR-based machine learning model to predict ICU sepsis and respiratory failure, enabling earlier interventions for veterans. Built production-ready ML pipelines, handling data preprocessing, model development, and evaluation.

Jul 2021 – Oct 2023 · 2 yrs 3 mos · United States

• Optimized the timing of patient chart reviews to increase the number of clinically appropriate suggestions shown on the review queue. Predicted whether a patient will have a diagnosis to add or drop, if reviewed by a reviewer. Worked with cloud engineers to build ML pipeline.

Aug 2018 – Aug 2021 · 3 yrs

• Applies the latest advances in machine learning and predictive analytics toward solving clinically and operationally significant prediction and classification problems across the medical center. Works on scalable infrastructure to more easily deploy and monitor machine learning based predictive analytic systems. Covers all facets of the machine learning pipeline, including model building, evaluation, deployment, operations, and maintenance/upgrade, as well as their integration with operational and decision-support systems at the hospital.

Jun 2016 – Aug 2021 · 5 yrs 2 mos

• Determining treatment algorithms for patient subgroups in stroke rehabilitation; developing clustering (biclustering and triclustering) algorithms to find patient categories associated with the success of a particular rehabilitation paradigm.
• Eye-hand coordination in real world reach-to-grasp; modeling index position and eye movement data for various shapes using linear mixed effects models and predicting the final index position using the gaze location and previous trial finger position in R.
• Power spectrum analysis of the muscle activation in professional pianists to determine symptomatic muscle activation patterns and predicting the level of impairment using linear mixed effects models in R.

Aug 2013 – Jun 2016 · 2 yrs 10 mos

Mar 2016 – Mar 2018 · 2 yrs · Greater New York City Area

• FitAI is a fitness-based IoT platform to provide real-time bio-mechanical weightlifting data to athletes using machine learning techniques.

Jun 2010 – Aug 2010 · 2 mos · Ankara, Turkey

• • Designed, simulated and implemented a 3dB Quadrative Coupler.