About me
I’m Komal. I build high-performance distributed systems, mostly in Rust.
I gravitate toward domains where latency and correctness matter — AI inference, real-time trading infrastructure, video processing at scale. The common thread is always systems depth: lock-free data structures, event-driven pipelines, bare-metal deployments, and getting the last millisecond out of the critical path.
In a room of ML engineers, I’m the one who can ship Rust production systems. In a room of Rust engineers, I’m the one who understands inference pipelines and on-chain settlement. In a room of crypto engineers, I’m the one thinking about systems-level performance optimization. The intersection is small.
Now
(Updated March 2026)
- Building and exploring interests - sandbox orchestration, trading systems, BFT consensus, RL envs, local ML…
- Based in Hyderabad, working remotely
Previously
Lead Engineer at YRAL (2024-2025). Short-video platform on the Internet Computer blockchain. High-throughput event-driven video processing pipelines, a hybrid Rust/Python recommendation system, infra migration, observability.
Recurse Center (2023). Built a distributed persistent KV store from scratch in Rust. Raft consensus, Bitcask storage engine, custom threadpool, async I/O. Also built a BitTorrent client, HTTP proxy, and SQL parser. Blog posts from the batch.
Co-founder at Cognirel (2019-2021). Pre-GPT conversational AI platform built on knowledge graphs. Led a team of 7. Acquired by VerSe/Dailyhunt.
Before that: Interview Kickstart (backend, async infrastructure, perf optimization), Flipkart (ML pipelines, image understanding), VMware. CS from BITS Pilani.
The long game
I care about intelligence and complex systems — artificial intelligence, financial markets, and eventually the nature of consciousness itself. These domains share more than people think: real-time decision-making under uncertainty, distributed coordination, and the challenge of building systems that are both fast and correct.
The decade-long arc points toward neuroscience, brain-computer interfaces, and consciousness research. The work I do now — whether it’s optimizing inference pipelines or building trading infrastructure — is all groundwork for understanding how intelligence works at the deepest level.
This blog
I write about systems internals, distributed systems design, and occasionally about neuroscience and how biological systems work. The 12 days of blogging series covers low-level topics like epoll, file systems, caching, and nginx internals.