Ben Kang Email & Phone Number

San Jose, California, Us

Recommendations

Menlo Park, Ca, Us

Large scale recommendations for Jobs and Newsfeed

New York, New York, Us

Tensorflow / TFX / Kubeflow / Recommendation System / Personalization / ML / AI

New York City, Ny, Us

• Productionize machine learning models and automate deployment process by using MLflow, Docker, Kubeflow, Kubernetes(EKS) in AWS environment• Developed end-to-end personalized recommendation system pipeline for NBC video streaming app by using Collaborative Filtering(Matrix Factorization, ALS - Spark MLlib) and Deep Learning(Tensorflow)

New York City, Ny, Us

• Designed and Performed A/B testing to experiment between multiple algorithms• Created ETL process by using Spark Streaming and Delta Lake

New York City, Ny, Us

• Built time series models using ARIMA and GAM to forecast daily/weekly business KPIs• Established time series analysis procedure in Python to process data from multiple sources and improve forecast accuracy• Created TV show scoring system and dashboard to deliver real-time performance of shows in digital platforms

Santa Monica, Ca, Us

• Designed collaborative-filtering recommendation model in Python(sklearn, tensorflow) to identify best candidates for clients

Ulsan, Kr

"Hyundai Heavy Industries Co., Ltd is the world's largest shipbuilding company providing 4 main business: Shipbuilding, Offshore & Engineering, Industrial Plant & Engineering, Engine & Machinery"• Automated ship design optimization by scripting in VBA to interface Excel and power prediction tool, helping the design team to reduce design variation M/H by 90% and improve ship efficiency by 3%• Developed ship initial design software add-ons, helping the Initial Design Team to reduce design appraisal document production M/H by 20%

• Developed software toolkit using C++ and Windows OS DLL hook API in Microsoft Visual Studio/SVN/ Redmine environment, automating field implementation procedure to reduce installation time by 20%

"Wing Ship Technology Corp. developed the world’s first commercial Wing-In-Ground (WIG) craft, a cutting-edge technology of marine transportation with 180km/h cruising speed and one-third of fuel consumption (US $44M)"• Synthesized and led cross-functional teams including internal and external stakeholders by showing proactive leadership and rigorous plan, receiving the craft’s safety certificate of Lloyd’s Register, U.K.• Designed data acquisition plan of the sea trial and developed the best flight model to forecast both safe and performance enhancing control parameters, achieving the craft’s successful first flight in 2013• Coordinated with international suppliers in more than ten countries, reducing half of the development cost but improving all functionalities of the craft to guarantee safety[WIG craft Propulsion System Development] 2012-2015• Project manager• Planned and organized the project. Wrote and defended US$1.3 million budget• Coordinated WIG craft application of aviation turboprop engine with Pratt & Whitney Canada(PWC)[WIG craft Certification of Lloyd’s Register(LR)] 2010-2013• Organized WIG craft approval plan with LR Design Appraisal Office in Rotterdam• Coordinated design and construction survey for local surveyors of LR• Developed and submitted relevant documents including "Structural plans and analysis reports", "Sea trial analysis reports" to LR

• Solved a mapping problem of different grid type computational models of 1 million nodes using search algorithm and OpenMP parallel computing API in C++, enabling engineering team to build realistic model• Automated transferring data between CATIA and Excel using VBA API to optimize design parameters and enable real-time visualization of design modification, reducing estimated design period to 25%• Developed motion simulation program by using C++ and numerical analysis to predict water resistance and impact load of the Wingship

• Researched the problem of craft’s speed stuck at 30 knots and solved by modifying the air flow in the air cushion system, increasing the ship’s speed to the designed take-off speed of 60 knots.• Analyzed structured and unstructured data from five different sources to build a structural model, predicting aerodynamic and hydrodynamic loads in 95% accuracy compare to the test flight results• Developed structural design(CATIA) and analysis for both airframe and ship hull(NASTRAN, PATRAN) by systemizing the design and test procedure• Initiated the Wingship production by contacting every approved manufacturers and coordinating them to provide customized parts and materials.

• Created a real-time ship behavior simulation solution using C++ to optimize and visualize the ship operation, helping to reduce the expected operation time by 70%

Master'S Degree, Business Analytics

Deep Reinforcement Learning Nanodegree, Computer Science

Bachelor'S Degree, Naval Architecture And Ocean Engineering

English