Xinglei Liu

I'm

About

Dedicated, enthusiastic M.S. student with passion for data science and engineering. Independent self-starter with proven ability to drive technical projects to successful completion through collaboration and execution of skills. Highly adept at collecting data, specifying requirements, building solutions, and conducting end-to-end analysis. Accomplished communicator with experience presenting methodologies and results to stakeholders at all levels.

Skills

Python 90%
SQL 90%
HTML 75%
Machine Learning 80%
Spark 75%
Java 65%

Resume

Professional Experience

Machine Learning Research Intern

2023 - Present

Tenvos Inc., West Sacramento, CA

Collected, processed, and analyzed audio data, performed feature engineering, and built machine learning models.

  • Extracted and evaluated acoustic features from 15,000+ diverse audio files, conducted statistical tests, and visualized results, enhancing model interpretability and delivering insights to customers
  • Built an automatic pipeline for audio data preprocessing, integrating transcription, speaker diarization, segmentation, and alignment, resulting in a 75% reduction in data processing time

Software Engineering Intern

2019

Yiheng Information Co. Ltd., Shanghai, China

Trained machine learning model and utilized Flask to create corresponding web API. Built regression model to predict stock prices of listed companies and designed algorithm to detect peaks and valleys.

  • Enhanced search process by building machine learning model to generate abbreviations from company names
  • Designed finance model to predict stock prices of listed companies and select promising stocks

Awards

ATHENAHACKS 2022 BEST HACK IN FINANCE

2022

Created a web application that helps tourists and LA residents to park easier with rates and time filters at an input destination.

  • Utilized real-time parking data and merged it with a fixed dataset of meter locations
  • Developed a user-friendly interface with a search bar to input destinations and utilized the Google Map API to convert them to coordinates

Education

Master of Science & Applied Data Science

2021 - 2023

University of Southern California, Los Angeles, CA

Bachelor of Science & Computer Science & Data Science

2017 - 2021

Purdue University, West Lafayette, IN

Academic Projects

NAME ENTITY RECOGNITION PROJECT

2023

Built a bidirectional LSTM model to recognize name entities with 90.8% F1 score.

  • Designed and implemented bidirectional LSTM model structure to predict NER tags and wrote code for training and validation using PyTorch
  • Achieved 20% improvement in model accuracy by optimizing hyperparameters, implementing cyclic learning rate scheduler, and using GloVe embeddings with boolean masks to handle case-insensitivity

ARTIST SEARCH WEB APPLICATION PROJECT

2022

Created a full-stack web application that allows its users to search for artists and explore artists details, artworks, and artwork genres.

  • Implemented backend using Node.js and Express framework
  • Created responsive, mobile-first frontend using Angular, TypeScript, and Bootstrap

JOB POSTING KNOWLEDGE GRAPH PROJECT

2022

Built a job search web application to facilitate our job seeking process by building connections between jobs and companies and gather as much diverse information.

  • Crawled 10,000+ pages from three different websites and improved crawling efficiency by multi-threading
  • Implemented rule-based information extraction using spaCy to and deduplicated entities using rltk

FARFETCH IMAGE CAPTIONING PROJECT

2021

Purdue University, West Lafayette, IN

Created ResNext machine learning model to generate labels with descriptions from fashion images, in partnership with client Farfetch. Interpreted data, evaluated models, and presented final model design and results.

  • Preprocessed and transformed image data and loaded to machine learning model for training
  • Identified two bugs and performed debugging steps in collaboration with team members
  • Implemented regular batch submission and checkpoint saving steps to reduce model training time

Contact

Call:

+1 7654763131