Focused representations
& topic discovery
FoTo connects targeted topic modeling with document visualization. My work on fSNE explores how to preserve local structure around selected points of interest.
FoTo · fSNE · Topic modeling
New Mexico State University
Sanuj Kumar.
Ph.D. Researcher · Computer Science
Natural Language Processing & Text Mining
I develop methods that help us find relevant information in large text collections and summarize it while preserving its meaning.
My research connects focused analysis, topic modeling, and unsupervised multi-document summarization, with applications in public-health communication.
Focused analysisRepresentation learningAbstractive summarizationTrustworthy NLP
01 / Research
What I work on
Methods for exploring, organizing, and synthesizing text around the information that matters to a user.
Focused representations
Meaning-preserving
summarization
I investigate unsupervised abstractive summarization across multiple documents and reviews, with an emphasis on preserving meaning, capturing consensus, and retaining query-relevant topics.
Ongoing doctoral research
Reliable NLP for
public-health messaging
With an interdisciplinary team, I develop human-in-the-loop methods to analyze public-health messages and distinguish contradictions, legitimate updates, and contextual differences.
NSF-funded PHMCS project
02 / Selected writing
Publications & preprints
2026
A Simulation-Based Slope Metric for Anchor List Reliability in Word Embedding Spaces
Sociological Methodology · 56(2), 147–176
2026
Selecting Language Models for Social Science: Start Small, Start Open, and Validate
arXiv:2601.10926 · Preprint
2024
FoTo: Targeted Visual Topic Modeling for Focused Analysis of Short Texts
LREC-COLING 2024 · Torino, Italy · pp. 7406–7416
2022
Focused Stochastic Neighbor Embedding for Better Preserving Points of Interest
IEEE/ACM BDCAT 2022 · REU Symposium · pp. 259–264
2021
A Word Embedding Topic Model for Robust Inference of Topics and Visualization
AIMLSystems 2021 · pp. 1–7
03 / Experience
Research & practice
2020–present
Research Assistant
New Mexico State University
NLP and text mining research spanning focused analysis, topic modeling, summarization, and interdisciplinary public-health communication.
2025–present
AI Research Fellow
Handshake AI
Evaluation and analysis of AI model reliability, reasoning, and failure modes.
2025–present
Instructor & AI Integration Reviewer
zyBooks · A Wiley Brand
Review of AI-enhanced computer science learning materials for technical accuracy and effective instruction.
Teaching & education
Learning, together
May–July 2025
AI/ML Trainer · Break Through Tech AI
Cornell University / Cornell Tech
Applied AI and machine learning instruction, hands-on activities, and support for learners building technical skills.
Summer 2024
Java Programming Instructor
CS 152/452 · New Mexico State University
Additional teaching experience includes support for Python, data structures, database systems, and introductory computer science.
Education
Ph.D. in Computer Science
New Mexico State University · In progress
M.S. in Computer Science
New Mexico State University · 2021
04 / Research in conversation
Selected presentations
October 29, 2025 · Stanford, California
FoTo at Stanford University
Poster presentation of FoTo: Targeted Visual Topic Modeling for Focused Analysis of Short Texts at the Center for Decoding the Universe Quarterly Forum, Stanford Data Science.
June 4–5, 2026 · Rochester, Minnesota
Public-health messaging at Mayo Clinic
Poster presentation of Human-in-the-Loop Agentic Auditing of COVID-19 Public-Health Messaging: Detecting Contradictions, Updates, and Differences at the Mayo Clinic AI Research Summit.
05 / Beyond research
Mentoring & community
Undergraduate research
REU & student mentoring
Supported undergraduate researchers with research computing, programming, experimental workflows, and the path from an idea to a research paper.
Graduate student leadership
Graduate Student Council
Served as Vice President of Activities, helping organize research events and opportunities for graduate students to share their work.
Student community
Indian Students Association
Former president and subsequent advisor, supporting cultural programming, student connections, and a welcoming community at NMSU.
Latest research · Preprint
Auditing public-health communication
Our latest manuscript on human-in-the-loop NLP for longitudinal analysis of evolving public-health communication.
August 2026 · JMIR Preprints
Auditing Evolving Public Health Communication with Human-in-the-Loop AI: Development and Validation of a Longitudinal Natural Language Processing Framework
Sanuj Kumar, Sumanth Reddy Nandhikonda, Heather Harper, Jagdish Khubchandani, and Marshall A. Taylor
Our Public Health Message Consistency System (PHMCS) combines NLP and human review to identify potentially inconsistent messages in evolving public-health communication.
Status: Submitted to JMIR Infodemiology; publicly available through JMIR Preprints.
Academic CV
Complete publication history, appointments, teaching, and service.
Preview CV on this page
If the PDF does not display on your device, use “Open CV” above.
Get in touch
Let's talk research.
For research collaborations, academic opportunities, or conversations about NLP and focused analysis.