Co-developed top-line mobile connectivity Metaverse readiness metrics, analyzed GPU cluster usage inefficiencies, developed roadmaps to improve internal partner GPU usage and provided insight into large-scale server fleet upgrade. • Analyzed GPU cluster usage for large-scale distributed ranking and content recommendation model training to find the root cause of efficiency bottlenecks. • As a tech lead, developed roadmap items in collaboration with software engineers and other data scientists to create automated systems to root cause efficiency bottlenecks faster. • Developed projects in collaboration with Reality Labs to improve GPU utilization in training neural network models for content understanding. Created the initial pipelines and dashboards to track GPU efficiency metrics customized for the Reality Labs. • Re-designed the monitoring dashboard for a company-wide server migration and supervised a contingent worker in the rebuilding of the dashboard. The dashboard became the central hub for product managers to plan and track the project and helped the migration of software pipelines on 4K inefficient servers to more efficient ones. • As a tech lead, developed research projects with software developers to analyze memory issues in ML training pipelines to reduce resource waste on GPU clusters and supervised junior members in building predictive models. • Co-led a group of data scientists in a technical training initiative to improve the onboarding experience of new hires for the Infrastructure Data Science organization. Facilitated the training of 60 data scientists and data engineers with customized content.

Samsung Electronics America | Mar 2021 - Jul 2021

Director, Data Science

Led a team of two engineers and a data scientist to maintain and extend the data science infrastructure capabilities of Samsung.com and create sustainable applied mathematics-based products. • Coordinated product development with internal customers, marketing departments and demand planners. • Developed products for demand forecasting and email marketing. • Trained junior team members. • Developed a Kanban board based project management system in collaboration with peers to prioritize and organize projects in the larger data science and engineering team.

Samsung Electronics America | Jun 2017 - Mar 2021

Senior Manager Strategic Analytics

Developed machine learning infrastructure and tools at Samsung.com using predictive and descriptive analytics for CRM marketing and web page personalization applications. • Led a team to build machine learning infrastructure to automate predictive model maintenance based on the champion challenger framework. • Developed predictive models to improve the effectiveness of marketing campaigns. • Developed an automation framework for demand forecasting at Samsung.com and related demand forecasting algorithms. • Lead a team to build a rapid classification model building tool based on AutoML ideas. • Lead a team to build tools based on contextual bandit algorithms and MILP for E-mail marketing.

Macy’s Inc. | Jan 2016 - Jul 2017

Data Scientist, Director

Built internal tool based on time-series analysis, optimization and predictive modeling to determine product demand in an omni-channel retail environment. • Build a to tool to predict buy quantities for internal merchants using Tableau, SQL and Python with direct report. The tool has been used across Macy’s to estimate demand for seasonal commodities. • Provided technical support for the roll out of the tool to the rest of the organization.

Collective | Aug 2013 - Dec 2015

Senior Data Scientist

Developed methods to deliver online ads more effectively across multiple platforms (desktop computers, laptops, cell phones and tablets) in a coordinated fashion using graph and ML algorithms. • Prototyped method to link cookies and advertising IDs on mobile devices in R and SQL based on graph algorithms and clustering techniques for use in daily production to process 500+ million cookies and identifiers. • Collaborated with software developers, product managers and business stakeholders to have the developed algorithms implemented in production. • Analyzed performance of the implemented methods in production and off- line. • Maintained production systems building hundreds of machine learning algorithms for predicting click-through rate and conversions.

Undertone | Feb 2011 - Aug 2013

Senior Engineer, Data Analysis and Algorithms Research

Developed algorithms to place online ads more effectively and validated performance in actual operation. • Worked on statistical models to predict click-through rate. • Computed possible profit improvements obtainable via better delivery algorithms using linear programming. Implemented the algorithm in C/C++ on actual server running code. • Provided statistical analyses to the operations team for better campaign management.

The Mathworks | Oct 2002 - Aug 2004

Real Time Simulation Technologies Developer

Programmed various software features for Simulink (a dynamic system simulator) and Real-Time Workshop (code generator for embedded processors) mainly in C/C++ and M (MATLAB). • Produced an enhanced debugger for Simulink as a member of a team of 4 engineers. Co-authored and was awarded 3 software patents. • Advised major clients on software setup and dynamic system simulation.

The Mathworks | Oct 2000 - Oct 2002

Simulink Engine Quality Engineer

Coded tests and built testing infrastructure in M-script and Perl to automatically test MATLAB, Simulink and related products. • Led a task force of 4 engineers to maintain and enhance the automatic testing infrastructure for memory and performance issues. • Tested and debugged C/C++ implementations to fix memory and performance issues. • As a member of a task force of 5 quality engineers, determined metrics to measure overall quality of the software products and assess readiness for release.

My Education

Take a look at my formal education

Massachusetts Institute of Technology

Boğaziçi Üniversitesi / Bogazici University

Udacity

Create Your Data Portfolio for Free