Software Engineer - Search & Indexing Platform, Apple Ads

Apple Ads helps developers and businesses reach customers on the App Store through highly relevant, privacy-focused advertising. Our engineering teams build the infrastructure that powers ad delivery at Apple scale - from search and retrieval to auction, ranking, and measurement.

Description

We are looking for a Software Engineer to join the team responsible for Apple Ads' core search and indexing platform - the system that retrieves the most relevant ad candidates in response to real-time search queries across our ad placements.

This platform operates at massive scale, handling tens of thousands of requests per second with strict latency requirements, and indexes millions of advertiser campaigns with complex targeting signals. We are investing heavily in evolving this platform: transitioning from batch-based architectures to streaming-first, near-real-time indexing, building modular retrieval infrastructure that scales efficiently across multiple ad placements, and introducing on-device ML scoring at retrieval time.

You will have the opportunity to work across the full search stack - from index construction pipelines and data ingestion, to query execution, candidate ranking, and production reliability.","responsibilities":"Design, build, and maintain large-scale distributed search and indexing systems that power real-time ad retrieval

Develop streaming index pipelines that reflect advertiser campaign changes in near-real-time, replacing legacy batch-only workflows

Build and optimize retrieval components including inverted indexes, approximate nearest neighbor (ANN/EBR) indexes, query tree generation, and candidate ranking at the leaf level

Implement ad eligibility and filtering logic at retrieval time - including keyword matching, audience targeting, budget enforcement, frequency capping, and policy compliance

Drive platform generalization - abstracting core retrieval capabilities (matching, eligibility, scoring, ranking) into a shared base stack that new ad placements can build on without rebuilding from scratch

Build auto-scaling infrastructure to eliminate manual capacity operations and support organic demand growth

Develop observability and debugging tooling to give engineers real-time visibility into index state, query behavior, and candidate retrieval outcomes

Contribute to offline experimentation infrastructure that allows ML teams to evaluate new ranking and matching models without risk to production traffic

Participate in on-call rotations and lead incident response for production search and indexing systems

Preferred Qualifications

Experience with embedding-based retrieval (EBR) or approximate nearest neighbor (ANN) search systems

Background in advertising technology - ad serving, candidate retrieval, auction systems, bid resolution, or L1/L2 ranking pipelines

Experience deploying ML models in latency-sensitive inference environments, including feature serving and model lifecycle management

Familiarity with A/B experimentation platforms and offline evaluation pipelines for ranking and retrieval

Experience designing and migrating systems from batch-only to streaming-first architectures

Track record of improving developer tooling and code quality in large, collaborative codebases - including test coverage automation, schema management, and CI/CD pipeline improvements

Minimum Qualifications

3+ years of software engineering experience with a focus on backend or distributed systems

Strong programming skills in one or more of: Rust, C++, Java, or Scala

Deep understanding of search and information retrieval fundamentals - inverted indexes, posting lists, forward/reverse index construction, tokenization, query parsing, and sharding

Hands-on experience building and operating high-throughput, low-latency distributed services with tight SLA requirements

Experience with streaming data infrastructure such as Apache Kafka, Amazon Kinesis, or equivalent systems

Familiarity with container orchestration (e.g., Kubernetes) and operating services in large-scale cloud environments