Making the “Net” Work in the Agentic Era Permalink
Talk, AI in Action: Conversations with UCSB Researchers, UCSB Library, Santa Barbara, CA
Talks
Annotated talks
Talks written up as slides with an edited transcript. Each page steps through the slides build by build, with what I said beside each one.
-
Making the “Net” Work in the Agentic EraAI in Action: Conversations with UCSB Researchers, UCSB Library · September 29, 2026 · 21 slides
How do we make secure and performant Internet access affordable for all? Data-driven policymaking, self-driving networks, and teaching for human-in-the-loop systems.
2026
Making the “Net” Work in the Agentic Era
Talk, "AI in Action" speaker series, UCSB Library, Santa Barbara, CA
A talk in the UCSB Library’s “AI in Action” series on how our group uses AI to accelerate networking research, inform broadband policy, and move toward networks that run themselves — and on the labor gap between research papers and the publicly accessible artifacts that agents are well suited to fill.
Informing Data-Driven Broadband Policy: Tools for Policymakers, Scholars, and Advocates
Tutorial, TPRC54 Pre-Conference Tutorial, American University Washington College of Law, Washington, DC
A three-hour pre-conference tutorial at TPRC54, presented with Laasya Koduru and Elizabeth Belding, introducing policymakers, scholars, and advocates to practical tools for measuring broadband availability, affordability, and adoption: the Broadband-Plan Querying Tool (BQT+), the platformized BQT-SaaS for public agencies, and an address-level approach to evaluating affordability policy.
Panelist (Automation; Data Generation panels)
Panel, Workshop on Generative AI in Networking, University of Chicago Paris Center, France
Agentic AI in Network Measurement and Infrastructure (moderated session)
Seminar, CAIDA AIMS Workshop, UC San Diego, La Jolla, CA
Making Networks Reliable with AI (Panel)
Panel, Google India Networking Research Summit, Bengaluru, India
2025
Making the Self-Driving “Net” Work: Developing Production-Ready ML Models for Self-Driving Networks
Keynote, INDIS Workshop, SC25 (Supercomputing), St. Louis, MO
Making the “Net” Work for All: From Measurements to Impact
Keynote, inaugural IMC Student Workshop, co-located with ACM IMC 2025,
Demystifying Network Foundation Models
Talk, Google Networking Research Summit, Sunnyvale, CA
Making the Self-Driving “Net” Work: Developing Production-Ready ML Models for Self-Driving Networks
Talk, 6GIC-CLICK Workshop on Telecom AI, Institute for Communication Systems, University of Surrey, UK
2024
In Search of a Networking Unicorn: Realizing Closed-Loop ML Pipeline for Networking
Seminar, SECDAI 2024,
In Search of a Networking Unicorn: Realizing Closed-Loop ML Pipeline for Networking
Keynote, 8th Keeping Networks Integrated Together (KNIT) Workshop, San Diego Supercomputer Center
Making the “Net” Work for All
Talk, AWS Networking Research Seminar; Georgia Tech; University of Chicago; Purdue University,
2023
Opening Remarks, ACM SIGCOMM 2023 Tutorial: Closed-Loop “ML for Networks” Pipelines
Talk, ACM SIGCOMM 2023 Tutorial, New York, NY
Closed-loop ML for Networking
Panel, Google Networking Research Summit (Panel), Sunnyvale, CA
In Search of a Networking Unicorn: Realizing Closed-loop ML Pipelines for Networking
Talk, IISER Bhopal; Stanford University,
In Search of netUnicorn: A “Thin Waist” Data-Collection Platform for Generalizable ML Models
Seminar, Intel MLWiNS Monthly Tech Talk,
RELIEF: RL-based Self-driving Wireless Network Management System for QoE Optimization
Seminar, MLWiNS Annual (2022), Mid-Year (2023), and Annual (2023) Workshops,
Democratizing the Development of Production-ready ML Artifacts for Networking
Talk, IIT Roorkee; IIT Delhi; IIIT Delhi, India
2022
Democratizing Networking Research in the Era of AI/ML
Talk, Google Networking Research Summit (virtual), virtual
AI/ML for Network Security: The Emperor has no Clothes
Talk, ETH Zurich, Switzerland
2021
Making the “Net” Work: Scalable, Flexible, and Intelligent Systems for Network Management
Keynote, STEERS Workshop (co-located with IEEE/ACM CCGRID), virtual