Shangyu Lai

Double major in Applied Mathematics and Computer Science at the University of Maryland, College Park.


Education

University of Maryland, College Park

Bachelor of Science
Computer Science
Applied Mathematics
August 2019 - Dec 2022

Experience

Math tutor

University of Maryland, Department of Mathematics

Moderating the section of teaching programming for mathematics projects . Assisting students to understand algorithms, to write pseudo code, to read documentation, and to search relevant information.

Sep 2021 - present

Independent Developer

Face‑recognition‑Based Authentication System

Utilizing a famous face recognition project, insightface, to authenticate users. Integrating API for easy adoption. Designed an API for easily adding/taking pictures and verifying user identity. Discovered bugs of the pytorch implementation of insightface. Contributed to the source code. Here is my fork. Creating a demo site of the system.

May 2021 - Sep 2021

Independent Developer

OCR‑Based Grading System

Jan 2021 - Aug 2021

Projects

Analysis of Covid and related policy data in the United States

University of Maryland, CMSC330

Used python, pandas, and numpy to analyze the data of Covid‑19 and related policy. Visualized data using plots and animation with geographical data. Full report.

Dec 2021

Alexa Skill: Professor Rating

Developed an Alexa skill called Professor Rating, which helps student know the professors they want to know. Utilized API of Alexa as well as AWS lambda. Some functions of the Alexa skills, such as that grabbing information from the internet is run by AWS lambda. Scraping data from multiple sites to get average professors rating and related comments. Tidied the data and write API for easy gathering information so that users can know the rating and comments from other students who took the professors’ course. Available in Alexa store.

Dec 2021

Skills

Programming Languages & Tools
  • c
  • Matlab
  • Ruby
  • C#
  • pytorch
  • numpy
  • pandas

Papers

TriView‑ParNet: Parallel Network for Hybrid Recognition of Applied in a project for a education company to automatically grade test paper.

Applied Intelligence (Accepted)

Skillfully implied multiple state‑of‑the‑art methods to recognize hybrid and overlapped strings. Designed the semi‑supervised learning module. Applied in a project for a education company to automatically grade test paper.

Apr 2021 - Aug 2022