James Verbus
I build AI systems for uncertain or adversarial environments.
After nearly a decade at LinkedIn where I was most recently as a Senior Staff Machine Learning Engineer building systems for anti-abuse, trust, and platform integrity, I chose to step away to work closer to the AI frontier.
My background began in rare-event physics: I earned my Ph.D. at Brown on LUX, one of the world's most sensitive dark matter detectors searching for faint signals deep underground.
These days I'm focused on agentic AI systems. For collaborations, talks, or hard problems worth comparing notes on, get in touch.
Featured Work
Open Source
Extended Isolation Forest for Distributed Spark/Scala Anomaly Detection
Extended Isolation Forest support for linkedin/isolation-forest, with random hyperplane splits, validation plots, benchmarks, reference parity checks, and edge-case tests.
Research
Finding AI-Generated Faces in the Wild
Research and engineering work on detecting AI-generated profile images in real-world settings, including CVPR workshop publication and LinkedIn Engineering write-up.
Workshop
Reinforcement Learning for Orbital Transfers
A hands-on Brown AI Winter School workshop connecting orbital mechanics, reinforcement learning, PPO agents, and practical model diagnostics.
Recent Writing
Aug 7, 2025
From Demos to Deployment: Insights from Berkeley’s Agentic AI Summit 2025
Agentic AI is graduating from cool demos to real‑world deployment. Hardware gains, open standards, rigorous evals, and secure governance are paving the way for billions of reliable, low‑cost digital assistants.
Feb 10, 2025
Exploring LLMs and RAG at the 2025 AI Winter School (Brown University)
Hands-on workshop on Large Language Models and Retrieval-Augmented Generation (RAG) for physics researchers. Using OpenAI API and LLaMA for domain-specific applications.
Sep 23, 2024
Announcing ONNX Support in Isolation Forest
LinkedIn's open-source isolation forest library now supports ONNX export, enabling deployment beyond Spark for streaming and edge inference applications.
Projects
Open-source and software work centered on linkedin/isolation-forest, a distributed Spark/Scala implementation for large-scale unsupervised anomaly detection.
Core Project
GitHub: linkedin/isolation-forest
I built and open sourced this distributed Scala/Spark isolation-forest implementation for large-scale unsupervised anomaly detection at LinkedIn.
Artifacts
Maven Central: com.linkedin:isolation-forest
Published artifacts for integrating the library into JVM-based data pipelines.
Artifacts
PyPI: isolation-forest-onnx
Python package for converting LinkedIn's isolation-forest model format into ONNX for portable inference.
Research / Publications
2024
Finding AI-Generated Faces in the Wild
G.J. Aniano Porcile, J. Gindi, S. Mundra, J.R. Verbus, H. Farid
CVPR Workshop on Media Forensics
Paper ·
arXiv:2311.08577
2023
Exposing GAN-Generated Profile Photos from Compact Embeddings
S. Mundra, G.J. Aniano Porcile, S. Marvaniya, J.R. Verbus, H. Farid
CVPR Workshop on Media Forensics
Paper
2024
Deep Learning to Detect Abusive Sequences of User Activity in Online Network
J.R. Verbus, B. Wang
US Patent 11,936,682
Patent
Talks / Videos
2026
Brown AI Winter School 2026: Reinforcement Learning for Orbital Transfers
A 2.5-hour interactive, hands-on workshop at the 2026 AI Winter School, hosted by the Center for the Fundamental Physics of the Universe at Brown University.
Venue/Host: AI Winter School 2026, Center for the Fundamental Physics of the Universe at Brown University.
2025
Brown AI Winter School 2025: Exploring LLMs and RAG
A 2.5 hour interactive workshop at the 2025 AI Winter School, hosted by the Center for the Fundamental Physics of the Universe at Brown University.
Venue/Host: AI Winter School 2025, Center for the Fundamental Physics of the Universe at Brown University.