Backend and infrastructure engineer focused on systems programming and production reliability. Currently a Member of Technical Staff at a YC-backed startup, building infrastructure for AI agents and durable user workspaces.
Recently built persistent sandbox infrastructure with directly mounted durable storage, reducing workspace setup from ~70s to ~10s and recovery from ~486s to ~14s while preserving user sessions across restarts and code updates. Also redesigned workflow synchronization, multi-client streaming, and CI test execution.
Selected projects:
vase — A Go pub/sub broker built on a lock-free ring buffer. 62x faster than Redis at 100 subscribers (111k vs. 1.8k msg/s), with zero message loss across 5 million messages at 1,000 subscribers.
iouring-wal — A write-ahead log built in Go directly on io_uring raw syscalls, without external libraries or CGo. 5.5x faster than naive fsync at 64 concurrent writers, with SIGKILL-tested crash recovery.
lsm-tree-go — An LSM-tree storage engine with bloom filters, sparse indexing, leveled compaction, and crash recovery. Profiled from 250k to ~2M writes/sec using pprof and achieved 5x BoltDB throughput on random writes.
Remote: Yes
Willing to relocate: Yes
Technologies: Go, distributed systems, storage engines, Linux internals, io_uring, concurrency, TCP, mmap, PostgreSQL, Redis, Kafka, systems programming
Résumé/CV: https://drive.google.com/file/d/1f9Hq8gj2uXhPOO5LHXJ9c6s8uJZ...
Email: ashishrathour1102@gmail.com
GitHub: https://github.com/AasheeshLikePanner
---
Backend and infrastructure engineer focused on systems programming and production reliability. Currently a Member of Technical Staff at a YC-backed startup, building infrastructure for AI agents and durable user workspaces.
Recently built persistent sandbox infrastructure with directly mounted durable storage, reducing workspace setup from ~70s to ~10s and recovery from ~486s to ~14s while preserving user sessions across restarts and code updates. Also redesigned workflow synchronization, multi-client streaming, and CI test execution.
Selected projects:
vase — A Go pub/sub broker built on a lock-free ring buffer. 62x faster than Redis at 100 subscribers (111k vs. 1.8k msg/s), with zero message loss across 5 million messages at 1,000 subscribers.
iouring-wal — A write-ahead log built in Go directly on io_uring raw syscalls, without external libraries or CGo. 5.5x faster than naive fsync at 64 concurrent writers, with SIGKILL-tested crash recovery.
lsm-tree-go — An LSM-tree storage engine with bloom filters, sparse indexing, leveled compaction, and crash recovery. Profiled from 250k to ~2M writes/sec using pprof and achieved 5x BoltDB throughput on random writes.
I write about the engineering behind my projects at https://medium.com/@ashishrathour1102.
Looking for backend, infrastructure, distributed-systems, storage-engine, or systems-programming roles.