The opportunity
At Affirm, we exist for the moments that matter—giving people a clear, predictable way to pay over time, with no hidden fees, no surprises, and no tradeoffs on what matters most.
What you'll do
Define and evolve workload safety standards for Affirm's EKS fleet, such as: CPU and memory settings, workload isolation and placement, tolerance for disruption, graceful termination to name a few
Own each standard end to end: research and prototyping, white-glove onboarding for early adopters, self-service guidance for everyone else, and automation and policy guardrails that keep every workload compliant from day one
Lead cross-team initiatives from design through fleet-wide rollout, mentor: other engineers, and set a high bar for design, code review, and operational readiness
Isolate workloads by criticality and resource profile so one misbehaving: workload cannot degrade critical paths such as checkout
Fix stuck workloads and node lifecycle problems, such as pods that ignore: termination signals, pod disruption budgets that block node cycling, and Karpenter timeouts that prevent bin packing and the removal of excess capacity
Determine the safe operating limits of each cluster (pod count, node count,: API server request volume), keep clusters within them, and develop control plane scaling patterns for bursty workloads such as scheduled batch jobs and large deployments
What they're looking for
- Analyze workload resource profiles across the fleet (CPU, memory, disk,: network) to guide EC2 instance type strategy, and automate that analysis as workloads change
- Measure each standard's impact on availability, latency, risk, and cost, and: use the results to prioritize the roadmap
- Participate in on-call for Kubernetes runtime issues across the fleet, and: turn recurring failure patterns into roadmap items
- Work closely with the teams that run services, batch jobs, and ML workloads: on EKS, and with the deployment, capacity, and SRE teams
