SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 243,290 - 295,250 / annual
Roblox operates a massive data platform processing petabytes of data daily to power analytics, ML, and product decisions across tens of millions of daily users. The Data Platform Compute team owns the compute layer: the Kubernetes (EKS) clusters where all Spark workloads run, and the software enabling engineers, analysts, and ML practitioners to submit and execute work reliably and cost-effectively.
As a Senior Software Engineer on this team, you will build and operate systems sitting between users and the cluster, with real ownership over problems where design is not yet settled. You'll help decide which technologies Roblox adopts next.
Key responsibilities include:
• Design, build, and operate core compute platform services including Spark job submission APIs, scheduling integration, and internal Spark libraries that users depend on.
• Deliver Spark Connect to support notebook workflows at scale, including provisioning and managing EKS compute with sensible isolation, resource limits, and cost controls.
• Build fleet management automation for EKS, such as blue-green rollouts that validate new AMIs and Kubernetes versions before reaching production, with minimal disruption to running jobs.
• Develop auto-tuning capabilities that analyze Spark history and event logs to recommend or apply job configuration changes improving performance and reducing cost.
• Operate and upgrade YuniKorn-based batch scheduling on Kubernetes, including queue design, quotas, fairness across tenants, and version upgrades.
• Evaluate emerging technologies such as native Spark acceleration (e.g., Apache Comet), run rigorous benchmarks on real Roblox workloads, and make data-backed recommendations.
• Define SLOs and metrics, build observability, and drive down toil, incidents, and compute spend across a multi-tenant platform.
• Partner across data engineering, ML, analytics, and infrastructure teams to understand needs, communicate changes, and migrate users to better paths.
The role is based in San Mateo with hybrid expectations: onsite Tuesday, Wednesday, and Thursday, with optional presence Monday and Friday.
REQUIREMENTS:
• 5+ years of experience building and operating production-grade distributed systems or infrastructure platforms.
• Hands-on experience running Apache Spark on Kubernetes (EKS, GKE, or self-managed) as a platform, including executor and driver behavior, resource tuning, and troubleshooting failed or slow jobs. Platform builder experience required, not just Spark user experience.
• Strong experience operating production Kubernetes clusters at scale, including upgrades, node lifecycle, autoscaling, and operators or controllers.
• Strong proficiency in at least one of Go, Java, Scala, or Clojure, with a track record of writing clean, well-tested, maintainable backend code. Python is a plus.
• Experience with AWS (EKS, EC2, IAM, S3) or GCP equivalents, and infrastructure-as-code.
• Experience taking a system from design through production operation, including on-call, incident response, and observability.
• Ability to influence design decisions across teams, mentor engineers, and drive alignment on cross-cutting changes.
• B.S. in Computer Science or equivalent practical experience.
BONUS QUALIFICATIONS:
• Experience with YuniKorn, Volcano, Kueue, or other batch schedulers on Kubernetes.
• Experience with Spark Connect, Livy, Kyuubi, or other interactive and gateway layers for Spark.
• Experience with Spark Operator, Karpenter, Helm, Terraform, or Argo.
• Experience building tooling using Spark event logs or metrics for tuning, cost attribution, or anomaly detection.
• Experience maintaining internal Spark builds or driving Spark version upgrades.
• Familiarity with native execution engines (Comet, Gluten, Velox) or JVM performance tuning.
• Experience migrating workloads from YARN/Hadoop to Kubernetes.
• Open-source contributions in the data or Kubernetes ecosystem.
• Experience operating at consumer-internet scale (100M+ users).
