Ti-Chung Cheng

Senior Applied Scientist, Microsoft (Jul 2026 – Present) #

Research Intern, Microsoft Research — Special Projects (Feb 2024 – May 2024) #

• Mentors: Madeleine Daepp, Robert Ness.
• Investigated empirically how large language models influence high-stakes decision-making via misinformation and generative propaganda.
• Analyzed 150K+ crowd-sourced articles using time-series, linguistic, and qualitative methods to quantify signal and trace propagation patterns.
• Contributed cultural and regional expertise to improve framing and validity of internal discussions and external publications (see related commentary: The Economist — “Video will kill the truth if monitoring doesn’t improve” (Mar 2024)).

Research Intern, Microsoft Research — Software Analysis & Intelligence (SAINTES) (May 2023 – Aug 2023) #

• Mentors: Denae Ford Robinson, Nicole Forsgren, Carmen Badea, Christian Bird, Tom Zimmermann, Rob DeLine.
• Designed and implemented GEMS, an agentic LLM system producing theory-driven metrics for team pairing using GitHub and DevOps signals. Built with GPT-4 API, AutoGen, Guidance, FLAML, and MySQL. (Project page)
• Introduced iterative prompt priming to elicit expert-informed metrics; evaluated via qualitative comparisons on DevOps performance proxies showing improved specificity and theoretical grounding.
• Defined a two-stage within-subject qualitative study to probe expert perceptions of LLM-assisted team-matching workflows.

Graduate Researcher, University of Illinois Urbana-Champaign (Aug 2018 – Jun 2026) #

• Designed, built, and evaluated a Quadratic Surveys system grounded in quadratic voting for preference elicitation in collective decision-making, using mixed methods (quantitative statistics + qualitative user research). Results presented at CHI / CI / CSCW venues. (Project page)
• Led multiple HCI projects on human–AI interaction in smart homes, spreadsheet analysis workflows, and design-process tooling; mentored students and coordinated experiments.
• Prototyped systems with React, Nest.js, and MongoDB; evaluated via interviews, surveys, click-stream logs, and in-lab behavioral experiments.
• Analyzed experimental data using open coding, thematic analysis, and Bayesian modeling to derive robust inference about preference intensity and collective insights.

Graduate Teaching Assistant, University of Illinois Urbana-Champaign (Aug 2018 – Jun 2026) #

• See Teaching for more details.

Software Engineer Intern, Salesforce — Lightning Component Services Team (May 2020 – Aug 2020) #

• Built a VS Code plugin in TypeScript reducing XML development time by ~50% for Salesforce engineers.
• Contributed to Red Hat XML VS Code extension (PR #292) and Salesforce VS Code extension (PR #2726).

Software Engineer Intern, Salesforce — Lightning Component Services Team (May 2019 – Aug 2019) #

• Built pipelines and designed three dashboards for front-end cache monitoring using Java, Grafana, and Splunk to visualize daily logs at billion-scale.
• Optimized dashboard queries by ~10× for readability and maintainability.

Machine Learning Research Intern, KKBOX (May 2018 – Aug 2018) #

• Mentor: Dr. Yian Chen
• Implemented an NLP pipeline for Mandarin named-entity recognition achieving >90% accuracy.
• Designed a pattern-based relation-extraction pipeline for cross-language music content using 3B+ music records.

Undergraduate Research Assistant, The Chinese University of Hong Kong (Dec 2015 – Dec 2017) #

• Mentor: Prof. James Cheng
• Conducted research on distributed hash-based nearest-neighbor search algorithms with a publication at SIGMOD.
• Built a 15% more time-efficient and scalable image-retrieval system compared to OpenCV FLANN Library.
• Implemented Collision Counting LSH in C++; details: Implementation and Analysis of Collision Counting LSH on Husky.

Web Administrator, Research and Information Hub (RIH), The Chinese University of Hong Kong (2016 – 2017) #

• Managed and maintained the RIH website infrastructure and content workflow for the department.
• Implemented site updates and automation scripts; coordinated with student groups and faculty for content publishing and events.
• Improved site reliability and reduced content publishing latency through simple CI scripts and optimized asset pipelines.

Summer Research Intern, National Center for High-Performance Computing (NCHC) (Jul 2015 – Aug 2015) #

• Mentor: Dr. Ann Hsing-Yen
• Designed and implemented a MySQL database pipeline to retrieve and analyze malicious fast-flux domains.
• Authored an intern report: Development of Bad Domain Tracking System.
• Contributed to follow-on research that used BDTS to detect IP changes of malicious domains over time (Tsai et al., TANET 2018).