Meta#

95 picked talks across 25 digests with a speaker from Meta, newest first.

RIPE 93#

• Multi-Layer Routing Domain Protection at Hyperscale: From Route Leak to Automated Enforcement · Core — Daria Shantalova, Express Backbone Team, Meta

@Scale: Product 2026#

• The future is products that grow themselves! · Core — Jubin Chheda (Meta)
• From On-Call to Always-On · Core — Sumeet Arora (Meta)
• Ship Fast, Break Nothing: How AI Tests What AI Builds · Core — Sophie Zeng (Meta), Sreeja Thummala (Meta)
• Preview Player Reliability and the Agentic Defense Loop on Edits · Core — Ravi Shah (Meta), Bhumi Sabarwal (Meta)
• Chasing a Moving Target: Going AI-Native on Edits' Video Seek Engine · Adjacent — Boris Baracaldo (Meta)
• Adopting AV1 for Real-Time Communication (RTC) at Scale · Adjacent — Yu-Chen (Eric) Sun (Meta)
• Flywheel of Media Models & Product Velocity · Adjacent — Rushaan Mahajan (Meta)
• Content Accessibility at Scale: AI Voice Translations at Meta · Adjacent — Esra Cansizoglu (Meta)
• Rhythm - Anticipatory Intelligence for Organizations · Adjacent — Sundararajan Subramanian (Meta)
• Fireside Chat: Designing for the Agentic Era · Wildcard — Michael Martinez (Meta), Alex Cornell (Meta)
• Live Panel: AI Generated Media Provenance at Scale · Wildcard — Steve Dilla (Meta), Mohit Diwan (Meta), Sai Chandramouli (Meta), Karthic Veera (Meta), Alexandre Mourachko (Meta)

PyTorch Conference 2026#

• Future of Distributed Communication in PyTorch: New APIs for Fault Tolerance, RDMA and Extensibility · Core — Tristan Rice, Kapil Sharma — Meta (PyTorch Distributed)
• Beyond Collectives: Building Fused Compute-Communication Kernels with PyTorch Symmetric Memory · Adjacent — Ke Wen (NVIDIA), Natalia Gimelshein (Meta), Kapil Sharma (Meta, PyTorch Distributed)
• MCCL: Fault-Tolerant Collective Communication for Large-Scale GPU Training · Adjacent — Ben Carver — Meta
• Monarch - Single-Machine DevX at Distributed Scale · Adjacent — Thomas Wang, Zach DeVito — Meta
• Precompile for Training · Wildcard — Bob Ren; Aaron Orenstein (Meta)

Capacity Europe 2026#

• Panel 2: Fibre, Routes & the Fragility of Global Connectivity · Core — Marc Halbfinger (Red HMH), Steve Roberts (EXA Infrastructure), Andy Palmer-Felgate (Submarine Cable Engineer, Meta), Paul Gabla (Alcatel Submarine Networks)
• Panel: Chokepoints & Shockwaves: Protecting the Cables Europe Can't Lose · Adjacent — Martin Prins (Dutch Subsea Cable Coalition), Ugne Kleinauskaite (Meta), Martin Dowling (CTO, Midgard Infra), Andy Bax (EXA Infrastructure)
• Panel: From FLAP to Frontier: Rethinking Europe's footprint for AI · Adjacent — Dawn Childs (nLighten), Ivo Ivanov (DE-CIX), Amar Khan (Meta), Falk Weinreich (Portus Data Centers), Giancarlo Giacomello (Aruba)

OCP Global Summit 2026#

• MetaRoCE: Rethinking RDMA for the AI presented by Meta · Core — Arvind Srinivasan (Meta), Sandeep Nagaraja (Meta), Balakrishnan Raman (AMD)
• ESUN: Open Ethernet for the AI Scale-Up Era · Core — Pratik Marolia (Microsoft), Manoj Wadekar (Meta)
• Tahan: An Open, Ethernet-Based Scale-Up Switch for Meta's MTIA Accelerators · Core — Ashutosh Grewal (Meta), Vimal Vasudevan (Meta Platforms inc.)
• Meta 102.4T Ethernet Switch · Core — Rongchun Zhou (Meta), Reza Niazmand (Meta)
• Integrated Optics at Scale for AI Infrastructure: From 51.2T CPO Validation to Meta's 102.4T Integrated Optics Switch · Core — Granthana Rangaswamy (Meta), Siamak Amiralizadeh (Meta), Xu Wang (Meta Platforms, Inc.)
• Operating Linear Pluggable Optics at Hyperscale: From Link Management to Production Reliability · Core — Amin Yekani (Meta), Qing Wang (Meta)
• Large-Scale Deployment of 800Gb/s Active Electrical Cables (AEC) in Meta Data Centers for AI Applications · Core — Jingyu Yu (Meta), Abhijit Chakravarty (Meta)
• Srv6 powered backbone networks · Core — Jasmeet Bagga (Facebook), Andrey Golovanov (Meta)
• Scale-across Networking: Panel discussion on routing and transport equipment implications · Core — Jeff Rahn (Meta), Yawei Yin (Microsoft), Xiguang - Henry Wu (Broadcom), Tad Hofmeister (Google)
• Achieving AI Cluster Reliability at Scale: Lessons in Reducing Workload Interruptions · Adjacent — Bob Wang (Meta), Ramachandra Reddy Kallam (Meta)

@Scale: Networking 2026#

• MetaRoCE: Meta's RDMA Transport · Core — Arvind Srinivasan (Meta), Kingshuk Mandal (Keysight Technologies)
• MetaRoCE: From Spec to NIC to Open Source · Core — Sandeep Nagaraja (Meta), Balakrishnan Raman (AMD)
• Fiber Networks Innovations for AI at Scale · Core — Fabrice Ouandji (Meta), Sebastian Gault (Meta)
• Physical Network Design at Scale + 3 min Q&A · Core — Richard Cziva (Meta), Brandon Premo (Meta)
• Co-Designing Communication for AI Accelerators · Core — Rajeev Nair (Meta), Wes Bland (Meta)
• Lightning Talk - The Plan That Learns · Core — Ben Vallis (Meta)
• Lightning Talk - AI4NI @Runtime · Core — YuLing Chen (Meta)
• Live Panel: Scale-Out Networking for AI Supercomputers · Adjacent — Rajiv Krishnamurthy (Meta), Joseph Provine (Meta), Andrew Dickinson (Oracle), Jim Julson (Coreweave), Jitendra Padhye (OpenAI)
• Lightning Talk - 10x Network Design Engineer · Adjacent — Maaz Mohiuddin (Meta)
• Live Panel: Building at the Roofline — Planning and Delivering Gigawatt-Scale AI Infrastructure · Wildcard — Omar Baldonado (Meta), Amrita Kenue (Meta), Lucas Benson (Microsoft), Nikita Parikh (Google)

NetUK3 (The Brewery)#

• Edge Network Evolution · Adjacent — Ben Ryall, Ian Cleary (Meta)

@Scale: Systems & Reliability 2026#

• Our Journey to Safely Unleash Agents at Meta Scale · Core — David Pariag (Meta)
• Extending Meta to the Public Cloud · Core — Dinesh Govindasamy (Meta), Sargun Dhillon (Meta)
• Teaching AI to Fight Fires: Autonomous Reliability Agents at Meta Scale · Core — Gaurav Mitra (Meta)
• Taming AI Infrastructure Failures with Agentic Debugging · Core — Phillip Liu (Meta)
• AI Storage Blueprint · Core — Sidharth Bajaj (Meta), Venkatraghavan Srinivasan (Meta)
• [Almost] Fail & Tell: Stop The World · Adjacent — Phil Lopreiato (Meta), Rahul Iyengar (Meta)
• Building Privacy Aware Infrastructure in the AI-Native Era · Wildcard — Rituraj Kirti (Meta)

@Scale: AI & Data 2026#

• How Meta Scaled AI Training Storage via Data Normalization · Core — Sarang Masti, Weiran Liu (Meta)
• Citrine: Catching Million-Dollar ML Anti-Patterns Before They Ship · Core — Prakash KL, Chenguang Zhu, Shyam Sundar Chandrasekaran, Jon Dyer (Meta)
• ASEA: Autonomous Agents for ML Serving Efficiency at Scale · Core — Bikash Sharma, Aashik Gowda, Angineh Keshishian, Mia Gu, Michael Fulthorp, Jinfu Leng (Meta)
• LP Planner: Synchronization-Aware Sharding Optimization for Large-Scale DLRM Model Training · Core — Mine Su Erturk, Greg Macnamara, Caner Gocmen, Isuru Janith Ranawaka, Felicity Liao, Ahmed Shuaibi, Alireza Tehrani, Hammad Ather, Shafeeq Ibraheem, Kaustubh Vartak, Nipun Gupta (Meta)
• Agentic Inference Optimization for RecSys ML models · Core — Shuyao Bi (Meta)
• SparkSentry: Autonomous Error Detection, Triage, and Repair for Spark at Scale · Core — Varun Srinivas (Meta)
• Data Governance in the World of Agents · Core — Komal Mangtani (Meta)
• Meta-ML: Supercharging ML Productivity at Scale with AI-Native Infrastructure · Adjacent — Niv Taiber, Bar Ulman, Ravali Busetty, Aasim Rab, Brant Swidler (Meta)
• XStream: Meta's unified stream processing service · Adjacent — Stella Kaval, Zheng Zhang, Shuo Xu, Jun Fan (Meta)
• Scaling Batch Graph Processing to Trillions of Edges · Adjacent — Sahil Gandhi, Arneish Prateek (Meta)
• Agent Memory at Meta Scale: One Query Language for Graph, Vector, and Memory · Adjacent — Abdullah Ozturk, Yiting Li, Vishal Gandhi, Ming Chen, Junjie Qi (Meta)
• Evaluating Opsmate: Grading AI Incident Investigations at Scale · Adjacent — Chinmay Gandhi, Narayanan Sankaran, Khushbu Thakur, Ankit Agarwal, Vaidyanathan PK, KC Balsu, Akash Jothi (Meta)
• Cache the Session, Not the Data: 40x Faster ML Iteration with a Shared Sidecar at Meta · Adjacent — Rohil Bansal, Eric Fu (Meta)
• Agentic Data Flywheel: A Multi-Agent System for Scalable Multimodal Training Data Production · Wildcard — Wendy Jiang, Selahattin Akkas, Alan Li, Antoine Simoulin (Meta)
• The Agent Feed: A Communication Layer for Multi-Agent Collaboration · Wildcard — Chinmay Gandhi, Vlad Tsvang, Patrick Walsh (Meta)

NANOG 97#

• Lessons From Scaling BPF To Detect RDMA Device Drivers Bugs In Real Time · Core — Prankur Gupta – Meta Platforms, Inc
• Network OS Upgrades at Hyperscale · Core — Sujitha Sekar Rajan – Meta Platforms Inc
• Crafting Network Configurations at Meta · Adjacent — Santiago Alfonso Munoz Rodriguez – Meta

MLSys 2026#

• MLCommons Chakra: Advancing Performance Benchmarking and Co-design using Standardized Execution Traces · Adjacent — Srinivas, Andrey Balogh, Brad B, et al. (MLCommons-led effort with industry participation from NVIDIA, AMD, Meta, Keysight, HPE, and others, per the paper)
• Optimizing Deployment Configurations for LLM Inference · Adjacent — Sungmin Cho, Jaewon Lee, Chunqiang Tang, et al. (Meta, serving the Llama family to ~1B monthly active users, per the paper)
• Sparing Strategies to Minimize Reliability Impact On Large Training Jobs · Adjacent — Kevin Quirk, Matthew Lennie, Ehsan K. Ardestani, et al. (Meta, per the paper's framing around Meta's AI clusters)

RIPE 92 (Edinburgh)#

• Evolution of Meta's Edge and Backbone Networks · Core — Fredrik Korsbäck, Meta Platforms, Inc

NSDI '26#

• Matryoshka: Realizing Hyperscale Data Center Network Design for the AI Era · Core — Yan Cai, Jialong Li (Max Planck Institute for Informatics), Kutalmis Akpinar, Tianxiang Li, Hany Morsy, Jason Wilson, Sunil Khaunte (Meta), Yiting Xia (MPI), Ying Zhang (Meta), et al.
• Express Lane to Efficiency and Reliability: Multi-Dimensional Control in Meta’s Express Backbone Network · Core — Faisal Iqbal, Vitaly Neganov, et al. (Meta), Aditya Akella (UT Austin), Ying Zhang (Meta)

MCP Dev Summit North America 2026#

• Bridging Kernel Space and AI: Building an MCP Server for Linux Scheduler Observability - Daniel Hodges, Meta · Core — Daniel Hodges, Software Engineer (Linux team), Meta

GTC 2026#

• Supercharging AI With Multi-Gigawatt AI Factories · Core — Gilad Shainer, Peter Salanki (CoreWeave), Omar Baldonado (Meta), Yuya Kawakami (SoftBank), Alexis Bjorlin (NVIDIA), Evan Burness (Microsoft) — Panel, Wednesday, March 18
• From Kernels to Clusters: How PyTorch Powers High-Performance AI (Presented by Meta) · Adjacent — Alban Desmaison (Meta) — Theater Talk, Monday, March 16
• Co-Engineering Grace CPU Platforms for Hyperscale · Adjacent — Shobhit Kanaujia, Abhishek Dhanotia (Meta), Ian Finder (NVIDIA) — Talk, Tuesday, March 17

NANOG 96#

• Keynote: AI Data Center Networking: Lessons from Meta's Evolution · Core — Omar Baldonado — Meta

FOSDEM 2026#

• Lessons from scaling BPF to detect RDMA Device Drivers Bugs in real time · Core — Prankur Gupta, Maksim Samoilov (Meta)

ACM CoNEXT 2025#

• The New (Pareto) Frontier of Cloud Routing: High Availability, Precise Control, or Stable — Choose Two · Core — Jiangchen Zhu; Ethan Katz-Bassett (Columbia University); Tom Koch (Google); Matt Calder (Meta); Italo Cunha (UFMG); Kevin Vermeulen (CNRS, École Polytechnique); Shuyue Yu; Ilgar Mammadov (Columbia University)

KubeCon + CloudNativeCon North America 2025#

• Fast and the Furious: CICD Pipeline for eBPF Programs at Meta Scale · Wildcard — Theophilus Benson (CMU) & Prankur Gupta, Meta — Wildcard

IETF 124 (Montreal)#

• SCONE TCP Option · Wildcard — Wesley Eddy, Matt Joras (Meta) — individual draft in TCPM

RIPE 91 (Bucharest)#

• PON in the Datacenter: Hyperscale for Management and Console · Core — Mike Joseph, Meta

OCP Global Summit 2025#

• Capacity Aware Adaptive Routing · Core — Speakers: Yikai Lin (Meta), Midhun Somasundaran (Meta)
• Evolution of Ethernet based switch platforms and fabrics to meet Meta's AI training clusters' scale-out AI fabric demands · Core — Speakers: Parvez Shaikh (Meta), Lingjun Wu (Meta)
• Evaluation of CPO Performance and Pluggable Optics Health for Reliable AI Infrastructure · Core — Speakers: Siamak Amiralizadeh (Meta), VIRAL LOWALEKAR (Meta)
• Development Challenges and Next-Generation Opportunities for Linear Drive Pluggable Optics (LPO) and Linear Receive Optics (LRO) · Core — Speakers: Nan Ge (Meta Platforms), Amin Yekani (Meta)
• INSPECT: Data analysis tool for Proactive Link Failure Detection on Meta's AI Infrastructure · Core — Speakers: Granthana Rangaswamy (Meta), Arushi Sharma (Meta)
• Accurate Cable Length Measurement for Optimized Data Center Networking · Core — Speakers: Jasmeet Bagga (Facebook), Mehak Mahajan (Broadcom)

Netnod Tech Meeting 2025#

• Meta – Big Trends and how it will change the Network · Adjacent — Janne Elvelid — META, Fredrik “Hugge” Korsbäck — META

SIGCOMM 2025#

• Falcon: A Reliable, Low Latency Hardware Transport · Core — Arjun Singhvi, Nandita Dukkipati et al. (Google, with Meta/Nvidia/Microsoft)
• Unlocking Superior Performance in Reconfigurable DCNs with Credit-Based Transport (Flare) · Core — Federico De Marchi et al. (MPI-INF, Meta, NVIDIA)
• Centralium: A Hybrid Route-Planning Framework for Large-Scale DC Network Migrations · Core — Yikai Lin et al. (Meta, UMass)
• PreTE: Traffic Engineering with Predictive Failures · Adjacent — Congcong Miao et al. (Tencent, MIT, UCSB, Meta, Tsinghua)
• Confucius: Intent-Driven Network Management with Multi-Agent LLMs · Adjacent — Zhaodong Wang et al. (Meta, JHU, Harvard, Stony Brook)

SREcon25 Americas#

• Lies Programmers Believe About Memory · Adjacent — Chris Down, Meta (Linux kernel memory management maintainer) — Adjacent