An Intimate Dinner
A private dinner for engineering leaders navigating cloud costs, GPU scaling, and the shift from reactive to predictive autoscaling.
The Premise
Cloud bills are climbing. AI workloads are exploding. And traditional CPU-based autoscaling can't keep up. Join a small group of VPs, Directors, and Heads of Infrastructure for an off-the-record conversation about what's next — hosted by the team that built KEDA, the open source autoscaler powering Microsoft, FedEx, Reddit, and Cisco.
What We'll Discuss
CPU and memory metrics react too late — after saturation, not before. We'll explore how event-driven signals like request rate, queue depth, and concurrency change the game for scaling decisions.
Every team is now running inference workloads, and most are over-provisioning GPUs by 40% or more. How do you right-size the most expensive resource in your stack?
Shifting the narrative from "we need more budget" to "we cut spend 30–40% and redirected it to product." Making infrastructure a strategic function, not a cost center.
Your Hosts
A founding maintainer of KEDA (Kubernetes Event-Driven Autoscaler) and former member of the Knative Technical Oversight Committee. One of the most respected voices in Kubernetes autoscaling, Zbynek is a frequent speaker at KubeCon and across the cloud-native ecosystem.
A founder and operator with deep roots in recruiting, startups, and venture. After years inside NYC’s startup ecosystem—as an engineer, recruiter, and VP at a leading early-stage VC fund—Brian has helped some of the best founders in the world build their earliest teams and go to market.
The Technology Behind Kedify Powers
KEDA is trusted by some of the world's best engineering teams
"We haven't touched our scaling config in months, and our bills dropped."
— Surag Mungekar, CISO, Rupert
Seating is limited to 12 guests. We'll confirm your spot within 48 hours.
No sales pitch. Just great food and real conversation.
