Senior Search Platform Infrastructure Engineer

The Autonomous Cloud group in the Platform Division builds and operates the foundational infrastructure that lets every engineering team at Mercari ship products safely and quickly - covering the Kubernetes platform, search infrastructure, observability, networking, and FinOps. Our charter is to make complex infrastructure simple to consume, so product teams can spend their time on customer value rather than on running systems.

The Search Infrastructure team - part of the Autonomous Cloud division - runs Mercari's search platform as a managed service for engineering teams across the company. We operate large-scale, product-critical Elasticsearch clusters and supporting systems (Redis, vector databases) that power product search across Mercari's marketplace. Our job is to abstract away the operational complexity of search so product teams can focus on building features. We are a small, deep team that combines hands-on operation of search clusters with in-house expertise on search-engine technology - partnering closely with internal Search and product teams.

Co-own the Elasticsearch cluster lifecycle: provisioning, sizing, upgrades, shard / replica strategy, and index management for product-critical search workloads.

Join the search infrastructure on-call rotation after ramp-up; lead incident response, root-cause analysis, and follow-through improvements.

Drive performance and reliability initiatives across the cluster, JVM, and query layers. Improve the search APIs and aggregation services consumed by product teams.

As bandwidth allows, collaborate with the Platform Infrastructure team on high-impact Kubernetes and Terraform changes as a senior reviewer and engineer; mentor engineers across both teams.

Operating a search platform that serves product-critical workloads at very large scale (hundreds of millions of indexed items at sustained high query volume) with demanding availability and latency expectations.

Owning the full search infrastructure stack end-to-end - from cluster operations and JVM tuning down to query-path performance - across a small, deep team where each engineer has real platform-level ownership.

Balancing day-to-day reliability of running clusters with longer-horizon investments: version upgrades, automation, and evaluating next-generation search and vector technologies.

Sitting at the boundary between infrastructure and product - translating between platform-level concerns (capacity, reliability, cost) and product-team needs (indexing patterns, query behavior, search experience).