hey i'm shubham, also kira. systems and backend engineer in jaipur, india , currently at commandcode on inference infrastructure and devops.
i work on gpus, model serving, deployments, apis, and the production paths connecting all of them. the parts i enjoy most are performance, failure modes, and tools people depend on when the easy path stops working. outside code i read and collect manga.
currently
software engineer at commandcode since june 2026
inference infrastructure, deployments, observability, and devops
high-throughput apis, gpu fleet reliability, and safe release paths
previously
co-founder & cto at routing ; stayed through acquisition and handover
senior software engineer at megallm
creator & maintainer of ultrabalancer
backend & devops at dyntech and groot music
open source
routing.run llm api gateway — routing, fallback, billing
ultrabalancer rust http/2 load balancer
sandbox disposable environments for humans and agents
airsec security review driven by the coding cli you already have
gopher ai discord bot with long-term memory, research, vision, and voice
tools
context bridge local-first handoffs between coding agents
sandbox disposable coding environments
watchman gpu inference monitoring
public edge
stack
community
gopher v2 an engineering-only server — no promo, no support queue, just people who build things
elsewhere