Shihao Ran

832-***-**** adbo0m@r.postjobfree.com Portfolio GitHub LinkedIn

Summary

Ph.D. student in Electrical and Computer Engineering pursuing a career as a Data Scientist .

Education

University of Houston

Houston, TX

Ph.D. Electrical and Computer Engineering

Aug. 2015 – Expected: May 2020

University of Electronic Science and Technology of China

Chengdu, China

Bachelor of Science, Electrical Engineering

Sep. 2011 – Jun. 2015

Skills

Programming

Python (primary), JavaScript

Data Analytics

SQL, MATLAB, R, Apache Spark, Tableau, Excel

Web/Media

HTML/CSS, Flask, D3.JS, Docker, Photoshop, AWS

Internship

Research Intern, Dataminr – AI Team, New York, NY

Sep. 2019 – Dec. 2019

Designed and implemented multiple data analytics and visualization strategies for end costumers and internal stakeholders based on the company’s product and internal data sets.

Deployed an AI-integrated timeline visualization platform (web-based) for retrospective analysis of large-scale events, taking care of both front-end (design, implementation) and back-end (model training and data processing).

Academic Projects

Deep Learning and Phase Retrieval in Holographic Imaging (CNN, ANN)

Aug. 2018 - Now

Developed a chemical holographic imaging system in the mid-infrared range for biomedical phase-sensitive imaging.

Optimized the ill-conditioned inverse problem using Keras-based Convolutional Neural Network (CNN) and Artificial Neural Network (ANN), achieved regression models with high-accuracy (>99.51%) and low noise sensitivity (<2.48%).

Published 4 research articles (with 14 citations) on Optics Express and IEEE journals.

Predicting Molecular Properties Competition (LightGBM)

Jul. 2019 – Sep. 2019

Built ensemble models (LightGBM) to predict the scalar coupling constant between pairs of atoms in a molecule.

Engineered more than 150 additional features based on the type of coupling and the structure of the molecule.

Implemented secondary feature engineering models for predicting meta-features. Decreased log MAE by 19.08%.

Real-time Twitter Analysis Web Application (Dash, Docker, AWS)

May 2019 - Jun. 2019

Developed a real-time analytics application on the cloud for Twitter data streaming and visualization (Python).

Implemented functionalities to track the most recent tweets for specific topics and display the changing trend of most-frequently-used words, including the corresponding sentiment analysis.

NLP: Automated Twitter User Tagging (SVM)

Apr. 2019 - May 2019

Built a Python pipeline to connect Twitter API and a PostgreSQL database for automated text mining and data migration.

Developed a binary Support Vector Machine (SVM) classier to identify and tag Twitter users by their profile description.

Benchmarked various models, including Logistic Regression, Random Forest, K-Nearest Neighbors, and Linear SVC for text classification. The SVM model achieved a 0.925 precision and a 0.929 recall.

Publications

S. Ran. et al. "Mitigating fringing in discrete frequency infrared imaging using time-delayed integration," Biomed. Opt. 832-843 (2018).

C. Wu, H. Le, S. Ran, M. Singh, I. V. Larina, and K. V. Larin, "Comparison and combination of rotational imaging optical coherence tomography and selective plane illumination microscopy for embryonic study," Biomed. Opt. (2017).

Contact this candidate