Xinyan Sun

• Conducted analysis on climate data to evaluate the probability of a forest fire;
• Ranked 7th among 53 groups; won “Best IoT Hack Using a Qualcomm Device” and “Wolfram Alpha Award”.

• Architected and developed an embedded voice assistant with Python and C on STM32 embedded platforms.
• Implemented cloud-based speech recognition; Conducted comprehensive evaluations for architecture selection.
• Led coordination efforts between software, hardware, and integration teams.
• Established benchmarking standards for client’s future projects; contributed code and insights.

• Designed and developed the back-end of a web platform, providing one-stop training for J&J new recruits, enabling task assignment, video streaming, resources sharing, and status tracking.
• Utilized Node.js and MongoDB for role based access control.
• Source code won over the work of another team, and was adopted by J&J for their internal use.

• Lead Design of an AI core with Python for a website to classify student answers and detect anomalies
• Utilized NLP encoders BERT and Universal Sentence Encoder for preprocessing, and a deep neural network with PyTorch for cheat detection
• Tested with ASAG Data Set and Real Exam Data with accuracy of around 80%

• Developed a web platform that pools and shares public or private music resources between playlist members to save on user efforts and costs.
• Implemented real-time collaborative decision making with Socket.io and JWT, and back-end with MEAN stack.
• Built ready-to-deploy Docker image with the prototype. Presented to the instructor and the class.

• Designed a web application that pairs up mentors and students with simple clicks on the phone.
• Developed front-end with Vue.js and Bootstrap; constructed server with Node.js framework Koa, and database with MongoDB; leveraged the idea of RESTful API.
• Deployed current version on web with Docker.

• Developed a website that optimizes course team partitioning with machine learning.
• Designed interactive survey creation and submission interface with JavaScript, CSS and HTML.
• Built back-end data storage with PHP and SQL.
• Deployed finalized version to the public with Google Cloud Platform, and adopted by the course professor to assign groups for the upcoming semester.

• Lead a team of three and developed a trading platform for RPI students;
• Read and stored items and user data from a MySQL database;
• Released the website on the university server.