Summary by AIEngineer.careers
Staff / Principal ML Infrastructure Engineer is a ML Infrastructure Engineer role (full-time). with Apple. in CUPERTINO, US. Imported listing (source: landoffer.ai). Apply on the employer's site (landoffer.ai).
Imported listing (source: landoffer.ai) · Apply on landoffer.ai
How our listings work
The sections below reproduce the third-party job description for reference. AIEngineer.careers does not write or control this text.
Imported job description
Sourced from landoffer.ai
Apple's AI Platform team is building an AI-native table format and high-throughput data-loading systems that power Apple Foundation Models and large-scale ML training. As a Staff/Principal ML Infrastructure Engineer, you will architect the on-disk format, indexing structures, transactional metadata layer, and Rust data plane and training data loader used by Apple's largest training workloads. The role spans petabyte-scale storage, distributed training data pipelines, and low-level I/O performance engineering across thousands of GPUs and TPUs.
Responsibilities
- Architect the on-disk table format, indexing structures, and transactional metadata layer
- Design and build the Rust data plane and training data loader for Apple's largest training workloads
- Engineer low-level I/O and memory performance (mmap, zero-copy, SIMD, async runtimes, object-store clients)
- Define transactional storage semantics including snapshot isolation, MVCC, and metadata scaling
- Set technical direction across teams and elevate senior engineers
Requirements
- 10+ years building storage engines, table/file formats, database internals, or high-performance distributed data and ML systems
- Proven track record designing table or file format internals (Iceberg, Lance, Delta Lake, Parquet, Arrow, RocksDB/LSM trees, or similar)
- Production systems programming mastery in Rust or C++ (Rust strongly preferred), plus Python/PyO3
- Deep expertise in low-level I/O and memory performance engineering (mmap, zero-copy, SIMD, tokio, S3/GCS client internals)
- Strong understanding of transactional storage semantics: snapshot isolation, MVCC, optimistic concurrency control, manifest trees
- Working knowledge of large-scale distributed training data consumption (JAX/Grain, PyTorch) across GPU and TPU meshes
- B.S., M.S., or Ph.D. in Computer Science, Computer Engineering, or equivalent practical experience
Nice to Have
- Core contributor or committer on Apache Iceberg, Apache Arrow, Parquet, Lance, Vortex, DataFusion, Velox, DuckDB, Polars, Daft, or RocksDB
- Experience scaling ML training data loaders and formats (PyGrain, ArrayRecord/Riegeli, Megatron-LM, Mosaic StreamingDataset, Ray Data, NVIDIA DALI/AIStore)
- Experience designing multimodal dataset storage
- Formal correctness or high-rigor verification practices (differential parity testing, property-based fuzzing, state-machine modeling)
Tech Stack
Apache ArrowApache IcebergC++DaftDataFusionDelta LakeDuckDBGCSGrainJAXLanceLSM treesMegatron-LMMVCCNVIDIA DALIParquetPolarsPyO3PythonPyTorchRay DataRocksDBRustS3SIMDTokioVeloxVortex

Cupertino · US · 223517+ employees
Apple is a global leader in designing, manufacturing, and marketing smartphones, personal computers, tablets, wearables, and accessories. The company also provides a variety of related digital services, software, and content platforms.