KAVANA VENKATESH / SELECTED ACADEMIC RECORD

Curriculum
Vitae.

A selective record of research, experience, teaching, and service, with brief descriptions and dates.

Google Scholar ↗Complete bibliography ↗

PROFILE / 2026

Ph.D. Researcher
Computer Science

Based at: Virginia Tech
Research: Agentic AI, multi-agent safety, collective behavior
Also: Generative modeling & image editing

RESEARCH / EDUCATION / SELECTED WORK / SERVICESCROLL TO READ ↓

01 / RESEARCH

Questions I work on.

I study how language-model agents coordinate, adapt, and sometimes fail together. My research focuses on safer multi-agent systems, emergent collective behavior, and reliable agent evaluation. I also work on generative vision and creative image modeling.

Multi-agent safety

Online monitoring, cascading failures, attribution, and intervention in LLM agent systems.

Collective intelligence

Coordination, emergent influence, and scalable agent-based modeling.

Generative models

Diffusion and flow-based methods for creative image generation and editing.

02 / EDUCATION

Academic training.

2024–present

Ph.D. in Computer Science and Applications

Virginia Tech Doctoral research with Dr. Jiaming Cui

2019–2021

M.S. in Data Science

Northeastern University Boston, Massachusetts

2014–2018

B.E. in Electrical and Electronics Engineering

Sri Jayachamarajendra College of Engineering (SJCE), India Undergraduate engineering

03 / SELECTED PUBLICATIONS

Selected research.

2026
NeurIPS 2026 Multi-agent safety
CASPIAN: Online Detection and Attribution of Cascade Attacks in LLM Multi-Agent Systems via Cross-Channel Causal Monitoring
2026
NeurIPS 2026 Generative vision
RAVEL: Rare Concept Generation and Editing via Graph-driven Relational Guidance
2026
NeurIPS 2026 · Agentic Web Workshop Collective intelligence
Do Agent Societies Develop Intellectual Elites? The Hidden Power Laws of Collective Cognition in LLM Multi-Agent Systems
2025
NeurIPS 2025 Generative vision
CREA: A Collaborative Multi-Agent Framework for Creative Image Editing and Generation
2025
CVPR 2025 Generative vision
FluxSpace: Disentangled Semantic Editing in Rectified Flow Transformers

View the complete publications catalogue

04 / EXPERIENCE

Research & practice.

Academic & research affiliations

Graduate Research Assistant

Virginia Tech

Research safe, reliable coordination in LLM multi-agent systems, including cascading failures and emergent collective behavior.
Develop evaluation frameworks, agent-based modeling methods, and generative vision systems.

2024–present

AI Safety Research Fellow

PRISM AI Safety Research Fellowship

Study decision-update consistency and model reliability under conflicting or evolving evidence.

2026–present

Selected industry experience

Machine Learning Research Intern

Apple

Researched user representations and machine learning for personalized content discovery.

Summer 2026

Applied Scientist Intern

Amazon Science

Worked on machine learning methods for planning and optimization at scale.

Summer 2025

Generative AI Data Scientist

iLink Digital

Built retrieval-augmented and LLM-powered applications for document intelligence.

2024

Data Scientist

Fidelity Investments

Applied NLP, computer vision, and machine learning to financial information workflows.

2022–2023

05 / TEACHING & SERVICE

Beyond the research.

Teaching

Virginia Tech2024–2025

Graduate Teaching Assistant: CS 4804 (Introduction to AI), CS 5805 (Machine Learning I), CS 4824 (Machine Learning).

Northeastern University2020–2021

Graduate Teaching Assistant: DS 2000 (Programming with Data).

Talks and teaching activities ↗

Peer review

ICML · NeurIPS · ACL · EMNLP · CVPR · ECCV · AAAI · TMLR · ACM TIST

Outstanding ReviewerCVPR 2026

Gold ReviewerICML 2026

Mentorship & outreach

PRISM Pathways Coach

Mentor aspiring and current Ph.D. researchers exploring AI safety.

2026–present

WiDS Worldwide Ambassador

Contribute to university workshops and outreach in data science and AI.

2024–present