Backend Engineering Intern

About Aight

Aight (Ashva Intelligence Pvt Ltd) is building India's AI infrastructure layer — a prepaid, INR-billed LLM gateway** built for Indian businesses. Think of us as the distribution-and-compliance layer for AI: GST-compliant invoicing, prepaid wallets with hard cutoffs.

Your code goes to production and touches real customer spend from week one.

What you'll actually work on

• The core gateway — a FastAPI service wrapping LiteLLM (as a library, not a proxy) to route requests across models, on a single invoice.
• Prepaid wallet + metering — real-time token accounting, hard balance cutoffs, usage estimation, and billing reconciliation. Getting the numbers right to the paisa is the job.
• INR / GST billing rails — the compliance layer that is our actual moat, not an afterthought.

What we're looking for

• Real backend fundamentals — you've built and shipped a REST API end-to-end. Python/FastAPI is ideal, but any stack where you genuinely understand the request lifecycle works.
• Real SQL — you can write a JOIN, know why an index matters, and can reason about a schema. PostgreSQL a plus.
• Debugging instinct — you chase the root cause, not the first Stack Overflow answer.
• Bonus: Redis, async Python, LLM APIs, or anything you've built with LiteLLM / vector DBs / RAG / voice.
• Bonus: you already build with AI coding agents (Claude Code, Cursor, Codex). It's how we work and you'll fit right in.

We care about what you've shipped, not where you studied. A deployed side project with real users, a hackathon win, or a real open-source PR beats a polished resume with an empty GitHub.

The details

• 6 months, full-time,
• in-office at our Gurugram officethis is a build-together, in-person role.
• Stipend ₹25,000–50,000/month, based on demonstrated ability.
• Strong performers get a full-time offer.