index
Impactian EToD™ Developer (2%)
Extraordinary Talent on Demand™
Kevin
Zhao
Python Developer
Software Developer in California, USA
Python
R
SQL
Matlab
Machine Learning
Deep Learning
Linear Programming
REST API
Docker
Flask
AWS
Linux
Bash
Git
HIRE
Python Developer
Impactian EToD™ Developer (2%)
Kevin Zhao
Software Developer in California, USA
Python
R
SQL
Matlab
Machine Learning
Deep Learning
Linear Programming
REST API
Docker
Flask
AWS
Linux
Bash
Git
HIRE
About
Kevin is a data scientist and engineer with several years of experience working in different industries, including but not limited to big data, energy, urban mobility, smart cities, and finance. He has worked at fast-paced startups building data products with full development cycles, including data engineering, machine learning, optimization, and data visualization.
Skills
Languages
Python, R, Matlab, SQL, HTML, CSS, JavaScript
Frameworks
Flask, Django
Libraries/APIs
Tensorflow, Keras, Scikit-learn, Pandas, Numpy, Scipy, Pyomo, Matplotlib, Bokeh, Plot.ly
Platforms
Linux, Docker, AWS, Google Cloud
Storage
PostgreSQL, MySQL, MS SQL, MongoDB, AWS DynamoDB, AWS S3
Other
Airflow, Apache Spark, AWS Lambda, AWS Athena, Cron, Metabase, Apache Superset
Tools
Git
Experience
Matlab
9 years
Python
6 years
SQL
5 years
R
3 years
Tensorflow
3 years
Keras
3 years
AWS
3 years
Flask
3 years
Scikit-learn
3 years
Highlight Projects
Nanogrid Software
Core software that connects to smart energy devices via REST API and optimizes energy usage to help homeowners build nanogrids with energy usage fully supported by solar + battery.
TouchLight Innovations
  • Connected to smart thermostats and other devices with 3rd party APIs.
  • Automated data pipeline using AWS EC2, Lambda, DynamoDB.
  • Built machine learning and deep learning models to predict indoor temperature and forecast solar production.
  • Optimized energy decisions with linear programming and deep learning.
Battery Health Classification
Internal tool to classify and recall batteries with manufacturing or performance defects from battery swapping energy network.
Gogoro
  • Queried time series-based battery swap data from PostgreSQL database.
  • Built and automated machine learning model to predict battery health with 92% recall.
  • Deployed machine learning model to Flask app using Docker.


Energy Demand Forecast
Time series forecast model to optimize operations of battery swapping network supporting over 200,000 electric scooter users.
Gogoro
  • Built time series forecast deep learning model with LSTM using Keras/Tensorflow.
  • Optimized model to perform with RMSE of less than 1 swap per station.
  • Deployed model to software platform with AWS.


Work Experience
CTO / Lead Data Scientist
TouchLight Innovations
|
Feb 2020 - Present
  • Lead the development for the core software product that managed building energy usage. This software helped homeowners build nano-grids with energy usage fully supported by solar + battery.
  • Built and deployed all machine learning, deep learning, and optimization models.
  • Collected data from 3rd party APIs and built data pipeline using cloud technology.
  • Participated in key business decisions to innovate the product.


Python
SQL
AWS EC2
AWS Lambda
AWS DynamoDB
Machine Learning
Deep Learning
Linear Programming
Flask
Data Scientist
Gogoro
|
Mar 2018 - Apr 2020
  • Developed energy algorithms for battery swapping network supporting over 200,000 electric scooter users and managing over 500 Mwh of energy.
  • Built Keras LSTM deep learning model to forecast demand of battery swapping network.
  • Developed anomaly detection model and designed experiments to predict internal short circuit in Li-ion battery packs.
  • Optimized charging algorithm to meet energy network demand and extend battery cycle life.
  • Built internal dashboards with open source tools to help R&D engineers diagnose problems within the energy network.


Python
SQL
Linux
Docker
Flask
Plot.ly
Machine Learning
Deep Learning
Airflow
AWS
Data Science Consultant
Independent Consulting
|
Feb 2017 - Dec 2018
  • FinTech Startup: Built and deployed supervised learning model with NLP to classify financial text data.
  • Green Building Startup: Analyzed and optimized ceramic processing data for sustainable building applications.


Python
R
Machine Learning
Deep Learning
MongoDB
AWS S3
AWS Sagemaker
Education
Northwestern University
Evanston, IL
|
Sep 2015 - Dec 2016
Master's Degree in Materials Science
Northwestern University
Evanston, IL
|
Sep 2011 - Jun 2015
Bachelor's Degree in Materials Science
About
Kevin is a data scientist and engineer with several years of experience working in different industries, including but not limited to big data, energy, urban mobility, smart cities, and finance. He has worked at fast-paced startups building data products with full development cycles, including data engineering, machine learning, optimization, and data visualization.
Skills
Languages
Python, R, Matlab, SQL, HTML, CSS, JavaScript
Frameworks
Flask, Django
Libraries/APIs
Tensorflow, Keras, Scikit-learn, Pandas, Numpy, Scipy, Pyomo, Matplotlib, Bokeh, Plot.ly
Platforms
Linux, Docker, AWS, Google Cloud
Storage
PostgreSQL, MySQL, MS SQL, MongoDB, AWS DynamoDB, AWS S3
Other
Airflow, Apache Spark, AWS Lambda, AWS Athena, Cron, Metabase, Apache Superset
Tools
Git
Experience
Matlab
9 years
Python
6 years
SQL
5 years
R
3 years
Tensorflow
3 years
Keras
3 years
AWS
3 years
Flask
3 years
Scikit-learn
3 years
Highlight Projects
Nanogrid Software
Core software that connects to smart energy devices via REST API and optimizes energy usage to help homeowners build nanogrids with energy usage fully supported by solar + battery.
TouchLight Innovations
  • Connected to smart thermostats and other devices with 3rd party APIs.
  • Automated data pipeline using AWS EC2, Lambda, DynamoDB.
  • Built machine learning and deep learning models to predict indoor temperature and forecast solar production.
  • Optimized energy decisions with linear programming and deep learning.
Battery Health Classification
Internal tool to classify and recall batteries with manufacturing or performance defects from battery swapping energy network.
Gogoro
  • Queried time series-based battery swap data from PostgreSQL database.
  • Built and automated machine learning model to predict battery health with 92% recall.
  • Deployed machine learning model to Flask app using Docker.


Energy Demand Forecast
Time series forecast model to optimize operations of battery swapping network supporting over 200,000 electric scooter users.
Gogoro
  • Built time series forecast deep learning model with LSTM using Keras/Tensorflow.
  • Optimized model to perform with RMSE of less than 1 swap per station.
  • Deployed model to software platform with AWS.


Work Experience
CTO / Lead Data Scientist
TouchLight Innovations | Feb 2020 - Present
  • Lead the development for the core software product that managed building energy usage. This software helped homeowners build nano-grids with energy usage fully supported by solar + battery.
  • Built and deployed all machine learning, deep learning, and optimization models.
  • Collected data from 3rd party APIs and built data pipeline using cloud technology.
  • Participated in key business decisions to innovate the product.


Python
SQL
AWS EC2
AWS Lambda
AWS DynamoDB
Machine Learning
Deep Learning
Linear Programming
Flask
Data Scientist
Gogoro | Mar 2018 - Apr 2020
  • Developed energy algorithms for battery swapping network supporting over 200,000 electric scooter users and managing over 500 Mwh of energy.
  • Built Keras LSTM deep learning model to forecast demand of battery swapping network.
  • Developed anomaly detection model and designed experiments to predict internal short circuit in Li-ion battery packs.
  • Optimized charging algorithm to meet energy network demand and extend battery cycle life.
  • Built internal dashboards with open source tools to help R&D engineers diagnose problems within the energy network.


Python
SQL
Linux
Docker
Flask
Plot.ly
Machine Learning
Deep Learning
Airflow
AWS
Data Science Consultant
Independent Consulting | Feb 2017 - Dec 2018
  • FinTech Startup: Built and deployed supervised learning model with NLP to classify financial text data.
  • Green Building Startup: Analyzed and optimized ceramic processing data for sustainable building applications.


Python
R
Machine Learning
Deep Learning
MongoDB
AWS S3
AWS Sagemaker
Education
Northwestern University
Evanston, IL | Sep 2015 - Dec 2016
Master's Degree in Materials Science
Northwestern University
Evanston, IL | Sep 2011 - Jun 2015
Bachelor's Degree in Materials Science