Engineer first. Curious always.
I'm Shivansh — a backend & AI systems engineer based in India (remote-first). I spend most days designing services that move money, messages, or embeddings between people and machines.
I started writing code because I wanted to understand how things stayed up. That's still the question I find most interesting: how does this system fail, and is the failure mode humane?
Outside of work I read systems papers, write occasionally about distributed primitives, and tune espresso variables like they're hyperparameters.
const profile: Engineer = { name: "Shivansh Shekhar", role: "Backend + AI", location: "India", focus: [ "distributed_systems", "llm_infrastructure", "developer_tools", ], availability: "remote", currentlyReading: "DDIA, 2nd ed.", };
I trace requests before I trace bugs.
If it pages someone, it isn't done.
LLMs are tools, not architecture.
Optimize for the codebase in two years.
Where I've worked
Designed and built enterprise Voice AI, logistics tracking, and unified payments backend architectures using Fastify v5, Turborepo, Hono, and PostgreSQL/Drizzle. Migrated IDP to Better Auth.
Developed distributed computer vision inference pipelines (Python/RabbitMQ), scaled Jasmin SMS gateway to handle 500k+ daily transactions, and built a Telegram support chatbot.