Justin Ho
AI product engineer and NLP researcher
justin-ho-1385
Brooklyn, New York, United States
Joined April 2026
Summary
Academic NLP and ML researcher: Justin has been active in academic research during his time at University of Florida and is a co-author on multiple 2025 publications (including FIDELITY and DETQUS), indicating a focus on natural language processing, summarization, and interpretable ML methods. aclanthology+2
Applied AI and product engineering: He holds product and engineering roles in industry (AI Product Engineer at GovSignals, Founding AI Engineer at Velocity Explorations) and has completed machine learning and software engineering internships, showing a blend of research and applied product work. rocketreach
Student-built tools and hands-on projects: As an undergraduate/graduate student he contributed to practical projects such as ZoomBuddy (Devpost) and participated in student organizations (The GAITOR Club presidency), reflecting experience building end-to-end student-facing software and leadership in campus groups. devpost+2
Early internship experience in engineering organizations: Multiple internships at Leonardo DRS (technical/software intern) and a 2023 machine learning internship at Ford indicate early professional experience in engineering and ML roles prior to or during graduate studies. rocketreach
Work
Education
Projects
Writing
FIDELITY: Fine-grained Interpretable Distillation for Effective Language Insights and Topic Yielding
April 1, 2025A hybrid method combining topic modeling and text summarization to produce fine-grained, semantically rich topic outputs and improve dataset interpretability; presented in Findings of NAACL 2025.
DETQUS: Decomposition-Enhanced Transformers for QUery-focused Summarization
March 1, 2025Introduces DETQUS, which uses tabular decomposition and a fine-tuned encoder-decoder model to improve query-focused tabular summarization; published on arXiv (Mar 2025).
Navigating the Blue Nowhere: A Framework for Mapping Validated Adversarial Trajectories
January 1, 2025Proposes the Cyber Behavior Pattern Extractor (CBPE), a neuro-symbolic framework to generate temporally-aware validated knowledge graphs from multi-modal cyber threat intelligence sources.