SUMMARY:

• A tech visionary who turns painful problems into profitable solutions at scale
• A deliberated learner who communicate what I learn by writing, coaching and building products. All happen by teaming up others.
• A people builder who nurtures and inspires you for your exponential transformation

TECHNICAL SKILLS:

Machine Learning (Regression, Classification, Clustering 8 years), Deep Learning (RNN, CNN, TensorFlow, Keras 2 years), Knowledge based AI (<1 year), Statistical Data Analysis (5 years), A/B Testing and Experimental Design (<1 year), Anomaly Detection (<1 year), Probability Graphical Model(1 year), Network Analysis (1 year), R(6 years), Python (numpy, pandas, plotly, sci - learn, nltk 3 years), MATLAB (8 years), CSS, HTML, JavaScript(<1 year), Business Analytics (1 year), GPU and High performance computing (<1 year), Hadoop (<1 year), Databricks & Spark(<1 year), SQLite, MySQL, NoSQL, MongoDB, PostgreSQL (1 year), Mathematical Modeling (8 years), Agent Based Modeling (4 years), Bioinformatics (5 years), Metabolomics (<1 year), Biochemical Systems Theory (8 years), Coaching and Mentoring (7 years), Project Management (2 years)

EXPERIENCE:

Confidential, Kansas City, MO

Data Scientist

Responsibilities:

• Developed three data driven software for eCommerce clients to reduce 1/3 operational cost, improve purchasing accuracy to 90%, and deliver personalized content.
• KPI: a ShinyApp and an R package using 7 key performance indicators to dynamically optimize pricing and purchasing under high degree uncertainty.
• OptimalSchedule: a ShinyApp and an R package dynamically optimize scheduling in an ever-changing environment; it utilizes one’s performance data to provide the best next steps with personalized instructions and recommendations.
• DonationAnalytics: a ShinyApp that describes and analyses the NGO's database of 1.2 million donation records over the spans of 15 years. It implemented five probability models to answer managerial questions such as * How many donors does the NGO still have? * How many donors will be active in one year from now? * How many transactions can be expected in next X weeks? * Which donors can be considered to have churned? * Which donors will provide the most value to the NGO going forward?
• Designed and conducted customer survey in an annual conference with 20,000 attendants. Used the survey to direct products purchase, to design educational program and to recommend products.

Confidential

AI Engineer and Data Scientist

Responsibilities:

• Built Automatic Speech Recognition engineer with reinforcement learning to coach kids of age 5-9 to read.
• Architected Python machine learning workflow on AWS
• Built language model and acoustic model adapted to kids’ speech by CMU Sphinx suite
• Used Google and AWS speech to text APIs
• Hand-engineered features important to identified kids' speech errors for ML

Confidential

AI Lead for Blockchain based Computing (Consultant)

Responsibilities:

• Responsible for optimizing dynamic computational networks based on blockchain using innovative Machine Learning algorithms and systems theory.
• Design multiple data science projects such as crypto currencies pricing prediction and recommendation, identification of computational sub-network for crypto mining, deep learning modeling on blockchain etc.

Confidential

AI Lead (Consultant)

Responsibilities:

• Led a team of 4 Stanford interns to build a deep learning classifier for the recognition of 7 common skin diseases with 92% accuracy.
• Led a team of 3 engineers to build a chat bot to guide people to make decisions in 5 areas to achieve optimal health.

Confidential

Cofounder & AI/ML Lead

Responsibilities:

• Head the design and development of social media platform for consulting & coaching based on technologies including AI, blockchain and AR.
• Built an NLP based classifier with 98.5% accuracy on testing data to recognize food borne illness comments for fast food companies, using NLP techniques to generate ML features and word embedding for deep learning models.
• Built an emojifier that turns input sentence into the most appropriate emoji, using deep learning techniques such as word embedding and LSTM.
• Built a speech recognizer that will activate desired app when trigger words are spoken, using deep learning techniques such as RNN and GNU.

Confidential, New Haven, CT

Postdoctoral Fellow

Responsibilities:

• Developed an integrative approach combining differential equations, machine-learning and bioinformatics to answer important questions concerning interferon stimulated genes (ISGs) expression dynamics and transcription regulation. The identified mechanisms provide novel targets for rational vaccine design.
• Established mixture regression methods to model data with large number of features, small sample size, and different origins. It selects important features for data from different subpopulations. The application to flu vaccination data shows significant improvement on predictive power over a single global model.

Confidential, Atlanta, GA

Graduate Researcher and Teaching Assistant

Responsibilities:

• Led five research projects in systems biology resulting in five first-author publications.
• Established integrative modeling frameworks to simulate and analyze hybrid biological systems consisting of interacting continuous, discrete, deterministic, stochastic components, which operate at different spatiotemporal scales.
• Created computational platforms that combine the deterministic modeling methodology of Biochemical Systems Theory with hybrid approaches, such as Petri Nets and Agent Based Modeling, being capable to integrate gene regulatory, metabolic and signaling pathways, for the analysis of diverse biological and biomedical data sets.