Jul 2018 – May 2020 · 1 yr 10 mos · Greater Chicago Area

• News analysis, Name Entity Recognition (NER), Research
• Build various Natural Language Processing (NLP) and Machine Learning models to improve document processing and understanding efficiency by > 90%. Projects/tooling include:
• Machine Translation (Transformer, BERT)
• Text Summarization (Pointer Network, Bi-LSTM RNN)
• Name Entity Extraction (NER)
• Document Classification (ELMo, GPT)
• Tooling: Spacy, Huggingface, PyTorch, Tensorflow, Keras
• Lead research of NLP development in KPMG Cognitive Lab. Enhanced multiple algorithms in the past 2 years by leveraging state-of-the-art (SOTA) solutions in the industry.
• Implement multiple NLP models to different requests from various clients (Financial Services, Technology, Professional Services, etc.) to help reduce labor costs significantly.
• Being an independent researcher and own full life-cycle of the products include research, prototyping, proof-of-concept (POC), testing, deployment, and production. Shared research results with team members from the dev team, data science team, and clients as a solid thought leader.

Nov 2017 – May 2018 · 6 mos · Ithaca, New York Area

• Led a team of 3 to design and draft marketing solutions by providing the client with data insights draw from customers' feedback data. Offer the client 4 viable marketing plans that will increase customer satisfaction, improve advertisement efficiency, and reduce cost significantly.
• Build Machine Learning, Deep Learning, and NLP techniques such as word2vec word-embedding, Gaussian Mixture Model (GMM), n-gram, and LDA topic modeling to identify customers' feedback data pattern.

Jun 2016 – Sep 2016 · 3 mos · Beijing City, China

• Developed a new algorithm independently to identify and filter possible paths of trash of Android apps. Reduce human efforts by > 90%.
• Experimented machine learning methods such as logistic regression and Bayesian model to identify possible paths
• Collaborated closely with cross-functional groups such as PM and Tech to improve my algorithm
• Analyzed different mobile apps and developed innovated business insights to assist executive team's decision-making

Jun 2015 – Sep 2015 · 3 mos · Greater Los Angeles Area

• Designed company webpage using online webpage designer Wufoo. Improve customer click-through-rate and completion rate by ~300%.
• Promoted company with marketing campaigns by analyzing web traffic data on Hubspot.

Sep 2014 – Oct 2014 · 1 mo · Hangzhou, Zhejiang, China