Software Engineer - Storage Systems (Agent Infra)

The Role

Storage decides how fast agent sandboxes start and how safely they share data. Agents need a filesystem that thousands of sandboxes mount at the same time, with strict isolation and the consistency agents need. Agent context is mostly small files, so the access patterns look very different from storage built for large files.

The Agent Infra team builds Agent Drive, the distributed filesystem that agent sandboxes mount over FUSE, and the storage underneath it. The work uses tiered caching, replication, content-addressed data, and Linux filesystem internals. In September 2026, Baseten acquired Blaxel whose sandboxes and Agent Drive filesystem are the foundation for Baseten Sandboxes

We are building a cloud, not just using one, so storage is designed for many tenants from day one. The team is small and ownership is big: we ship in weeks and months, not multi-year programs.

If you care about page cache coherency, FUSE, replication protocols, and content-addressable storage, we'd love to talk.

Example Initiatives

• Blaxel is joining Baseten to build the future of agentic infrastructure
• Baseten Sandboxes documentation

Responsibilities

• Design and implement core Agent Drive features: metadata replication, volume placement, garbage collection, and FUSE client behavior
• Solve multi-attach correctness: cache coherency across many FUSE mounts of the same drive, dentry and inode invalidation, write ordering, and crash consistency
• Optimize tail latency on read and write paths, including page cache strategy, writeback, prefetch, and metadata RPC batching
• Build tenant isolation into the storage layer, including quota enforcement, blast-radius containment, and noisy-neighbor mitigation
• Own the storage side of sandbox startup and resume performance
• Build durable block storage for sandboxes alongside other Baseten storage teams
• Build observability into every layer, drive storage incidents end to end, and turn each one into a permanent fix
• Contribute upstream to the open-source storage projects we build on

Requirements

• Strong systems programming background in Go, Rust, and/or C/C++, with comfort reading kernel code
• Deep understanding of Linux filesystem and storage internals: VFS, page cache, dentry and inode caches, FUSE, the block I/O stack, fsync semantics, and crash consistency
• Hands-on experience building distributed storage systems, content-addressable storage, or large-scale caching infrastructure
• Experience profiling I/O, memory, and RPC bottlenecks and optimizing tail latency in production
• Ability to debug correctness bugs that only show up under concurrency or partial failure
• You move fast and like owning a problem end to end on a small team

Nice to Have

• Experience with a distributed filesystem or object store such as JuiceFS, CephFS, or S3-compatible storage, or contributions to those projects
• Knowledge of consensus and replication protocols (Raft, leader leases, quorum reads)
• Experience with snapshot, clone, and copy-on-write filesystem features
• Experience operating multi-tenant storage with hard isolation requirements
• Familiarity with Linux performance tooling such as eBPF, io_uring, perf, ftrace, and blktrace

Benefits

• Competitive compensation, including meaningful equity
• (U.S. only) 100% coverage of medical, dental, and vision insurance for employee and dependents
• Flexible PTO policy including company wide Winter Break (our offices are closed from Christmas Eve to New Year's Day!)
• Paid parental leave
• Fertility and family-building stipend through Carrot
• (U.S. only) Company-facilitated 401(k)
• Exposure to a variety of ML startups, offering unparalleled learning and networking opportunities.

Apply now to embark on a rewarding journey in shaping the future of AI! If you are a motivated individual with a passion for machine learning and a desire to be part of a collaborative and forward-thinking team, we would love to hear from you.