LATEST POST IS UP →
systems, rust, and infrastructure notes

Work

I help teams solve hard infrastructure problems in Rust — whether that's scaling an existing system, fixing a performance bottleneck, or building something new from the ground up.

Notification Engine

  • 3.3B push + 70M pull notifications/day
  • 300k notifications/sec · 99.999% uptime
  • Android, iOS, Web — decoupled dispatchers

Performance Engineering

  • $1M/yr infrastructure cost reduction
  • Python hot paths rewrite in Rust
  • Custom reverse proxy replacing managed load balancers

Graph Database

  • 1M+ nodes, 4M+ edges, fully in-memory
  • Subgraph queries at p99 < 4ms
  • 4M edge writes in 1.4s, lock-free

Zero-Downtime Migrations

  • Zero-downtime Redis migration via CDC
  • CPU-aware throttling to preserve runtime stability
  • Resumable checkpointing for safe recovery

Stream Analytics

  • 3.3M events/sec throughput
  • 1B events analyzed in 5 min on a single 4-core/16GB node
  • Real-time rolling aggregates via DBSP

ML Feature Store

  • 60M daily requests at p99 < 5ms
  • 24 months zero downtime
  • 600k+ hot DB reloads, zero faults