Overview
Engineer and Technical Lead,, Dropbox Facundo is currently the tech lead for Dropbox's exabyte-scale blob storage system. These teams manage everything from customized storage machines with many petabytes of capacity to the client APIs other teams use internally. Prior to working at Dropbox, Facundo worked at Amazon on a variety of scheduling problems for Amazon's last mile delivery team. Ecosystem Enabling Director, Datacenter - AI Business, Intel Vice Chair, Board of Directors, SNIA Richelle Ahlvers runs Technology Initiatives and Ecosystem Enabling for the Datacenter and AI business for Intel, promoting and driving enablement of new technologies and standards strategies.
Speakers
12
Facundo AgrielStaff Software Engineer and Technical Lead,, Dropbox
Gaurav AgarwalDistinguished Engineer, Marvell
Jens Birk AndersenPhD Industrial Researcher, Samsung Electronics- Brian AtkinsonHPC Engineer, Los Alamos National Laboratory
- Fellow, AMDFellow, AMD, Fellow, AMD
- Fellow, EverspinFellow, Everspin, Fellow, Everspin
- Fellow, Hewlett Packard EnterpriseFellow, Hewlett Packard Enterprise, Fellow, Hewlett Packard Enterprise
- Mohannad Fahim AliI.O Protocol Architect, Intel
- PhD Student, Virginia TechPhD Student, Virginia Tech, PhD Student, Virginia Tech
- Research Scientist, Western DigitalResearch Scientist, Western Digital, Research Scientist, Western Digital
- Richelle AhlversEcosystem Enabling Director, Datacenter - AI Business, Intel
- Vivek Shah
Schedule
7:30 AM - 8:30 AM Registration Open - Breakfast
Breakfast available
Keynote 8:15 AM - 8:30 AM Welcome to SNIA Developer Conference 2026 J Metz, Chair, SNIA Board of Directors / Technical Director for Systems Design, AMD
Data Movement & Placement 8:30 AM - 9:20 AM Toward GPU‑Agnostic Storage: RDMA-Accelerated Data Movement Across Heterogeneous GPU Systems Jason Goldschmidt, Distinguished Engineer Dell
Keynote 8:30 AM - 8:50 AM StorageAI - 1 Year of Progress Jason Duquette, Co-Chair, SNIA StorageAI Community; Distinguished Engineer, Chief Platform Architect Dell Technologies
Data / Storage Architecture 8:30 AM - 9:20 AM To DPU or Not to DPU… Where Does the Work Belong? Jonathan Hinkle, Senior Director of Storage Pathfinding Solidigm
Breakouts
Welcome and Keynotes
Data Movement & Placement 9:30 AM - 10:20 AM FDP: The Data Placement Promise of Modern NVMe SSDs Vivek Shah, Staff Engineer Samsung Electronics Hui Qi, Staff Engineer Samsung Electronics
Data / Storage Architecture 9:30 AM - 10:20 AM Emulating a Full RDMA NIC in Userspace with VFIO-User: Architecture, Backends, and Lessons from Reaching 2 GB/s RDMA Without Hardware Stephen Bates, Fellow AMD
9:50 AM - 10:05 AM Break
Keynote 10:05 AM - 10:45 AM Optimizing HDD Performance in Modern Object Storage Systems Abhay Kataria, Senior VP, Software Engineering for Devices and Systems, Seagate Technology
Keynotes and SNIA Awards Presentation
10:20 AM - 10:35 AM Break
Morning Break
Data Movement & Placement 10:35 AM - 11:25 AM Characterizing and Emulating FDP SSDs with WARP Inho Song, Graduate Student Virginia Tech
Keynote 10:45 AM - 11:25 AM Lessons From the Front Lines of Enterprise AI: Real World Impact, Infrastructure Realities, and What Engineers Must Understand Rebecca Weekly, Chief Technology Officer brk tech, A Berkshire Hathaway Group Company
Keynote 11:25 AM - 11:45 AM SNIA Awards Presentation J Metz, Chair, SNIA Board of Directors / Technical Director for Systems Design, AMD
Data Movement & Placement 11:35 AM - 12:25 PM Concepts for HDD Data-Placement Optimization for Data Lakes Curtis Stevens, Technologist Seagate Technology
11:45 AM - 12:30 PM Lunch
12:25 PM - 1:10 PM Lunch
Lunch
Data / Storage Architecture 12:30 PM - 1:20 PM Opinionated Experts Provide their NI towards AI on Bandwidth, Interconnect and Cooling Enhancements Anthony Constantine, Distinguished Member of Technical Staff, Micron Technology John Geldman, Director, SSD Industry Standards KIOXIA Trent Johnson, Hardware Architect, IBM / Technical Council, SNIA IBM Glen Hanna, Principal System Architect Lenovo, ISG
12:30 PM - 1:30 PM Lunch
AI - General 1:10 PM - 2:00 PM Prompt Engineering for Storage Developers: From NVMe Drivers to Protocol Compliance Testing George Pamboris, Senior Director NetApp
Data / Storage Architecture 1:30 PM - 2:20 PM Scaling Data, Scaling the Ecosystem for GPU-initiated Storage Access CJ Newburn, Distinguished Engineer, NVIDIA Vikram Mailthody, Senior Researcher, NVIDIA
Data / Storage Architecture 1:30 PM - 2:20 PM Maximizing E3.L 2T: Engineering Ultra‑Dense QLC SSDs Trent Johnson, Hardware Architect, IBM / Technical Council, SNIA IBM
Data Movement & Placement 2:10 PM - 3:00 PM GPU-Initiated NVMe: Cutting the CPU Out of the Accelerator-to-Storage Data Path Stephen Bates, Fellow AMD
2:20 PM - 2:35 PM Break
Data / Storage Architecture 2:35 PM - 3:25 PM AI Impact on Storage Mike Klemm, Fellow, Technology and Storage Pathfinding, KIOXIA America
3:00 PM - 3:10 PM Break
Afternoon Break
File Systems & Protocols 3:10 PM - 4:00 PM Beyond POSIX: Building a Protocol-Native VFS for User-Space NAS Ben Jarvis, Principal Architect DDN
3:20 PM - 3:35 PM Break
Toward GPU‑Agnostic Storage: RDMA-Accelerated Data Movement Across Heterogeneous GPU Systems
This session presents a practical approach to GPU-agnostic storage using RDMA-enabled data paths to move data directly between disaggregated storage and GPU memory across NVIDIA, AMD, and Intel platforms.
Attacking the Core Problems of Data Access Control
Current data security measures are demonstrably not working in practice. Data breaches continue to occur because real systems are routinely deployed with misconfigurations that allow unauthorized access to data.
Closing the NVMe-CPCS Stack Gap: An End-to-End Implementation and Per-Workload Break-Even Characterization
The growth of AI and large-scale analytics workloads has made data movement a primary bottleneck of modern storage systems.
Data Storage Innovations for Scalable AI Infrastructure: From NAND Flash to Exabyte-Scale Hierarchies
As AI models grow from billions to trillions of parameters, storage has emerged as a fundamental bottleneck — not just an afterthought — in AI system design.
FDP: The Data Placement Promise of Modern NVMe SSDs
NVMe Flexible Data Placement (FDP) is rapidly emerging in commercial SSDs and early hyperscale deployments as the next evolution of host visible SSD placement control.
Who Encrypted My Volume? Pinpointing Ransomware VMs with Storage-Side Coherence Scoring
When ransomware attacks a shared storage volume in a virtualized data center, the storage array may detect signs of trouble - rising entropy, random-write bursts, and abnormal write IOPS - but it usually cannot identify which VM or host is responsible.
Leveraging CMB to Offload Memory Bandwidth for RDMA Direct Applications
High-performance SSDs use onboard DRAM for FTL tables and internal management with DRAM capacities increasing as SSD densities grow.
Storage for AI 104 - A Continuing Intro to Storage for Inference
Building on Storage for AI 103, which introduced why Inference needs KV Cache, this presentation will go deeper into the key types of storage used in Inference.
Characterizing and Emulating FDP SSDs with WARP
Flexible Data Placement (FDP) is the NVMe interface that hyperscalers such as Google and Meta have championed to reduce write amplification.
Living the Post Quantum Crypto Dream
There has been much discussion about quantum-resilient cryptography (QRC), with a heavy focus on algorithms and the shift to integration.
Computation Offload Specification for AI Frameworks to Emit Partitioning Functions Alongside Large Data-centric Tasks
After surveying large-scale AI and HPC workloads, we have identified a set of principles and best practices widely used in partitioning large sparse tensors and graphs.