Sr.Software Dev Engineer, DDB Index and Search

Senior Software Development Engineer, DynamoDB Search and Indexing Amazon DynamoDB is looking for a Senior Software Development Engineer to help build two new initiatives: native vector, full-text, and hybrid search; and Krupa, which aims to provide strongly consistent global secondary indexes (GSIs). These initiatives expand how customers access operational data—from finding semantically relevant results to querying current data through alternate keys. You will work on problems at the intersection of database internals, distributed systems, information retrieval, and consistency protocols. This role offers the opportunity to shape foundational architecture, write performance-critical software, and lead engineering work across multiple DynamoDB components. Search introduces a new set of indexing and retrieval challenges. How do you organize high-dimensional vectors across a partitioned dataset? Combine approximate nearest-neighbor retrieval with selective attribute filters? Rank results that mix semantic similarity and keyword relevance? You will help build the infrastructure behind these capabilities, including vector and inverted indexes, distributed query execution, and hybrid ranking. Your designs will balance recall, relevance, latency, freshness, and cost while the underlying data changes continuously. Krupa addresses a different foundational problem: strongly consistent access through GSIs. Asynchronous index propagation creates a gap between a successful table write and its visibility in an index. Closing that gap requires reasoning about updates across independently partitioned table and index data, concurrent operations, transaction interactions, and partial failures. You will help design and implement the mechanisms that preserve these guarantees through retries, recovery, and changes in partition ownership. Across these initiatives, the engineering challenge extends through the entire index lifecycle. Indexes must be built over existing data while writes continue, maintained as workloads evolve, and recovered when machines fail. You will work closely with engineers across storage, replication, transactions, routing, and the control plane to make these capabilities correct, performant, and operable. As an SDE III, you will own significant technical decisions and carry them through implementation, validation, launch, and production operation. You will lead designs spanning component boundaries, identify risks early, and create interfaces and implementation plans that enable other engineers to deliver independently. This is an opportunity to combine deep systems engineering with technical leadership on new DynamoDB capabilities. Key job responsibilities - Own architecture and implementation. Lead substantial work across search and indexing, translating ambiguous requirements into precise designs, working software, and measurable customer outcomes. - Build vector and text retrieval infrastructure. Develop vector indexing, approximate nearest-neighbor search, inverted indexes, text analysis, and relevance scoring. - Deliver distributed search execution.Design partition selection, query fan-out, filter execution, candidate retrieval, top-k merging, and hybrid ranking using techniques such as rank fusion or reranking. - Develop strongly consistent indexing mechanisms. Design and implement coordination between table mutations and index visibility, including concurrent updates, index-key changes, retries, and failure recovery. - Build the index lifecycle. Deliver online creation, backfill, incremental maintenance, compaction, rebalancing, and deletion while customer traffic continues. - Make guarantees testable. Define correctness invariants and search-quality criteria; build concurrency tests, failure-injection scenarios, and benchmarks that expose architectural weaknesses. - Optimize the complete system. Investigate CPU, memory, storage, and network bottlenecks; quantify tradeoffs among retrieval quality, coordination overhead, write amplification, throughput, and tail latency. - Lead across teams. Resolve design questions with partner teams, define clear contracts, and drive changes that span storage, replication, transactions, and query processing. - Raise engineering capability. Mentor engineers, lead design and code reviews, and break complex work into independently deliverable pieces with clear testing and rollout plans. - Own production outcomes. Build diagnostics and recovery mechanisms, participate in operations, and use production evidence to improve architecture and reliability.