Tin (Kevin) Nguyen
I am a PhD student in Computer Science at
Auburn University, working with
Prof. Anh Totti Nguyen and Prof. Chirag Agarwal at UVA. I work on Explainable AI, and Human-AI Interaction. I am particularly interested in improving human-AI workflows, and measuring frontier AI performance.
Publications my favorites () | patents () | others ()
Accepted as a Poster
PageGuide is a browser extension that acts as an in-page navigation agent: given a user's question, it locates the supporting evidence on the current webpage (DOM or non-DOM elements) and highlights exactly where the answer comes from, instead of leaving the user to skim the whole page.
TMLR, 2026
Also accepted at NeurIPS 2025 Workshop Multimodal Algorithmic Reasoning (MAR) —
Oral
Highlighted Chain of Thought (HoT) outputs XML-style evidence tags in each reasoning step, linking conclusions to specific input spans so users can trace why an answer is produced.
NAACL, 2024
PEEB is a part-based, interpretable-by-design image classifier that predicts an object's class by first describing its visible parts in natural language (e.g., wing color, beak shape) and matching those descriptions against a text-conditioned part detector. As a result, users can inspect exactly which part attributes drove a prediction and edit the class descriptions to correct or adapt the model without retraining.
Knowledge-based Systems (KBS), Jul 4, 2023
This work studies vision-and-language navigation in 3D environments built on VizDoom, where an agent must follow a natural-language instruction to reach a target location.
ICML, 2026
I contributed two tasks evaluating object counting and physical reasoning capabilities of video-language models.
International Journal of Advanced Computer Science and Applications (IJACSA), Vol. 11, No. 4, 2020
Korean Patent No. 10-2979119 — Registered Jun 2026
Expert Systems with Applications, 2023
Teaching
- Teaching Assistant for Formal Languages, Machine Learning, and Security at Auburn University
- Teaching Robotics for kids at K-6 AI Club at Auburn University
Mentoring
- Truong Thoi Thang, CS PhD student at Auburn University, 2026-Now
Academic Service
- Area Chair for Interpreting Agent Behavior (IAB) (NeurIPS 2026)
- Program Committee (PC) for IUI 27 (Intelligent User Interfaces)
- Reviewer for ML Conferences: ARR Rolling Review (ACL, EMNLP, NAACL), Evaluation of Interactive Agents (NeurIPS 2026), Interpreting Agent Behavior (IAB) (NeurIPS 2026)
- Reviewer for HCI Conferences: UIST, HAI, IUI
- Reviewer for Journals: Knowledge-based Systems
Education
Auburn University
Ph.D. in Computer Science and Engineering
Advisor: Anh (Totti) Nguyen
Research: Explainable AI and Human-AI Interaction
Teaching Assistant: Formal Languages, Machine Learning, Secure Software Process
Sejong University
M.Sc. in Computer Science and Engineering
Advisor: Kim Yong-Guk
Research: Robot Navigation
University of Science
B.A. in Computer Science and Engineering
Advisor: Ngoc Quoc Ly
Research: Computer Vision and Robotics
Work Experience
Aikyam Lab, University of Virginia
Research Intern
Advisor: Chirag Agarwal
- Developed a multimodal web agent that browses websites and captures evidence (text, images, charts) at every navigation step
- Developed a defense against model distillation attacks, in which an adversary copies a proprietary model by training on its outputs
AIOZ Co., Ltd.
AI Engineer
- Built three video-analysis systems: automatically reading car license plates, tracking a moving ball, and recognizing the same person across different cameras
Gumi Co., Ltd.
AI Engineer Intern
- Built an optical character recognition (OCR) model for handwritten Japanese text in scanned bills and documents
Gameloft
Game Programmer Intern
- Built an arcade game using Cocos2d-x (C++)
Honors & Awards
Tinker Research Grant
Tinker · Awarded $5,000 for compute/API credits
Korean Patent Registration ★ Patent
Ministry of Intellectual Property, Republic of Korea · Patent No. 10-2979119, Coarse-to-Fine Fusion Method and Apparatus for Language-Command-Based Virtual Space Navigation
GSC Travel Fellowship
Auburn University Graduate School Council · Awarded $400 to support conference travel
Tinker Promotion
Tinker · Awarded $150 in credits
1st Prize, AICovidVN Challenge
Ho Chi Minh, Vietnam · Introduced Fruit-CoV, a framework detecting SARS-CoV-2 through cough sounds (~$4,300 prize)
Skills
Programming Languages
Python
C/C++
HTML/CSS
JavaScript
TypeScript
R
AI & ML
LLMs
VLMs
AI Agents
RAG
Browser Harness
MCP
MCP
Frameworks & Libraries
PyTorch
TensorFlow
Hugging Face
vLLM
LangChain
LangGraph
Tinker
Weights & Biases
FastAPI
React Native
Node.js
Next.js
Robotics
ROS
Databases & Vector Stores
PostgreSQL
SQLite
Supabase
Redis
Neo4j
Infrastructure & DevOps
Docker
AWS
Azure
GCP
CI/CD
RabbitMQ
Developer Tools
Vim / Neovim
WezTerm
tmuxProjects
Resources
Links and course materials I keep coming back to.
Job Application Tips
Practical notes summarized from Karen Kelsky’s The Professor Is In, focused on presenting an academic record clearly, specifically, and with evidence.
Writing Strong Job Documents
- Lead with facts and evidence rather than emotion, enthusiasm, or unsupported claims.
- Edit aggressively: remove lists, repetition, filler, and details that do not serve the specific application.
- Write as a future colleague—not as a supplicant—and get repeated feedback from experienced readers.
Writing a Strong Cover Letter
- Use professional formatting, normal business-letter conventions, and approximately two pages for an early-career application.
- Show evidence of research, teaching, publications, future plans, and collegial contributions instead of making generic claims.
- Move quickly from the dissertation to its broader contribution, publications, second project, teaching, and specific departmental connections.
- Tailor every letter to the job, department, programs, faculty, and campus; avoid flattery and vague “fit” statements.
Writing a Strong CV
- Prioritize readability: consistent 12-point type, one-inch margins, clear headings, left-aligned entries, and visible years.
- Use a CV—not a résumé: avoid bullet points, columns, boxes, narrative paragraphs, personal stories, and descriptions of routine duties.
- Organize content around peer-reviewed and competitive achievements: education, appointments, publications, awards, grants, talks, teaching, research, service, skills, and references.
- List accomplishments in reverse chronological order and follow the conventions of your discipline and country.
Writing a Strong Teaching Statement
- Keep it concise—typically one page—and replace a life story with a clear teaching principle supported by evidence.
- Use specific courses, assignments, methods, and outcomes; show what students did and what they learned.
- Avoid clichés, excessive emotion, humility, and claims such as “I am passionate about teaching.”
- Adapt the emphasis to the institution, connect teaching to research, and end with a strong statement of broader educational purpose.
Presenting Evidence of Teaching Effectiveness
- Choose relevant, teachable syllabi with clear descriptions, appropriate workloads, assignments, and policies.
- Make proposed courses appealing to students, innovative, connected to the position, and complementary to existing offerings.
- List courses, levels, and enrollment briefly; summarize evaluations rather than sending an overwhelming archive.
- Prioritize evidence from courses where you were the primary instructor over routine TA materials.
Writing a Strong Research Statement
- Usually keep it to about two pages in the humanities and soft social sciences; follow field-specific norms elsewhere.
- Center the research: explain the overarching agenda, dissertation argument, methods, contribution, publications, and future work.
- Include a distinct second project so the record shows continuity, independence, and a sustainable research trajectory.
- Use confident, active language; minimize “interest” talk, grand claims, negativity, unnecessary citations, and references to other application documents.
Writing a Diversity Statement
- Connect awareness of inequity and access to concrete choices in teaching, research, mentoring, and service.
- Use specific experiences, practices, and outcomes rather than identity claims or vague commitments.
- Show how you work with students from different backgrounds and prepare them to participate in a diverse society.
Writing a Dissertation Abstract
- Frame the big disciplinary problem, briefly position the literature, identify a gap, and present your project as the response.
- Summarize the material, theory, methods, chapter contributions, original argument, and wider significance.
- Keep the research and contribution central; avoid process narratives, jargon, literature-review overload, and empty adjectives.
Learning Vim
- I became a keyboard warrior (Luysru) A YouTube video about going keyboard-first with Vim.
- Vim Adventures A browser game that teaches Vim movements and commands as you play.
- Vim Hero Interactive, hands-on lessons for learning Vim in the browser.
- Vimwise My own web app for learning Vim, inspired by Vim Hero.
Statistics in HCI
- Yatani: Statistical Methods in HCI A practical guide to the statistical tests commonly used in HCI research, with examples in R.
- How to Report an F-Statistic Scott MacKenzie's short note on reporting ANOVA results correctly.
- Penn State STAT 502: ANOVA and Design of Experiments Full course notes on ANOVA and experimental design.
Human-Robot Interaction
- Bajcsy: Human-Robot Interaction Course materials and notes on the algorithmic foundations of human-robot interaction.
JdeRobot Robotics Academy
- JdeRobot: Robotics Academy Exercises Hands-on robotics exercises (perception, navigation, driving, drones) built on ROS/Gazebo.