Duke University

Computer Vision at Overjet

Greater Boston

Overjet

Computer Vision

Leading body keypoints effort in addition to continuing the object detection work for cabin monitoring scenarios for production automotive programs

Affectiva

Senior Computer Vision Scientist

Lead object detection efforts running on low powered embedded SoCs for clients in the automotive space

Computer Vision Scientist

Built predictive models using video data for clients for a diverse set of tasks including depression analysis and ad skip modeling
Transitioned facial expression and action recognition classifiers from SVM-based methods to deep learning methods

Data Scientist

Performed network modeling and resilience
of Hurricane Sandy's impact on NYC MTA using public data and simulated different network recovery strategies.

risQ

Data Science Consultant

I am the Data Science lead in an effort to scale and automate data analysis currently done at PowerAdvocate. This involves data analysis and algorithm prototyping in Python and then assisting in the implementation of these models in SPARK and MLLib.

I have
-  Applied unsupervised and supervised machine learning techniques on Spend and Material Master Data in order to automate the classification process
-  Performed Basic NLP analysis using Topic Models, Word Embeddings, Named Entity Recognition, Regex etc...
-  Developed analysis tools using Python, MySql, HBase, Django, Scala
-  Used Python libraries including Pandas, Numpy, Bokeh, pymysql, Scipy, sklearn, and gensim
-  Created and assisted in production of metrics and testing to classify Spend data
-  Created an ETL (extract, transform load) pipeline and aggregated previously silo'ed Spend Data
-  Built a simple Django/ AngularJS App

PowerAdvocate

Worked on a quantitative model for sales prediction given personnel, organizational, and technological factors

Assisted in development and creation of the Client Advisor and future Partner program as well as various client engagements

Conducted business analysis of Revenade, provided short term and long term strategy as well as operations recommendations now being implemented

Implemented recommendations such as standardizing client engagements, the standardization of prospecting and programming of software for mass prospecting (market value greater than $50,000),  and expansion of services through offering analysis of existing sales data using big data partnerships

Revenade

Associate Consultant

Programmed an alkalinity data processing routine for automated alkalinity measurement of
brine, groundwater, and waste-water samples to be used in place of previous colorimetric
titration method in VBA.

Recorded an instructional video on the new automated method and computer routine
designed for standard usage in oil and waste-water industries for alkalinity measurement.

Created a project schedule that allowed for the collection of over 50 samples, titrations,
testing, programming, writing, and video recording within three months.

Kevin Wang

Kevin Wang