About Anuvaya
We think conversational AI agents will deliver all professional services in India. We started with astrology. We're a small group of engineers, designers, and product folks building at the intersection of conversational AI and domain expertise. Making an AI agent sound human-like is hard. Making an AI an expert in a domain is also hard. We're doing both together.
We're backed by Accel, Arkam Ventures, and Weekend Fund.
The Role
Most AI agents treat conversations as stateless request-response cycles. Ours doesn't. We've built a stateful agent that streams responses with natural pacing, handles mid-response interruptions without corrupting state, decides autonomously whether to keep talking or wait, and maintains a living context that evolves across turns.
The Agent Orchestration team owns the core runtime - the system that coordinates LLM calls, tool execution, context management, delivery pacing, and conversation persistence. When you break something here, users notice in real-time. When you get it right, the agent feels like talking to a person who genuinely knows what they're doing.
What You'll Do
- Own and evolve the stateful agent runtime - the streaming pipeline, state management, and tool execution layer
- Ensure what the user sees, what the model remembers, and what we persist stay perfectly in sync - even when users interrupt mid-response
- Build and refine the streaming architecture - how responses are buffered, delivered, and recovered cleanly on interruption
- Implement and tune delivery pacing - making the agent's output feel natural, not robotic
- Design the tool execution layer - the agent calls domain-specific tools mid-conversation and reasons over results in real-time
- Work on autonomous continuation - the agent deciding when to keep going and when to stop, based on conversational context
- Optimize for latency and reliability - first-token time, interruption recovery, context window management
What We're Looking For
- Deep systems programming instincts - you think about state machines, concurrency, race conditions, and failure modes naturally
- Experience building real-time or streaming systems - WebSockets, SSE, or similar
- You've built and maintained stateful, long-running processes in production
- Comfort with distributed systems - message queues, event-driven architectures, service coordination
- Strong debugging skills for systems where 'it worked on my machine' is never the problem
- You can read and reason about concurrent code - whether it's Elixir processes, Go goroutines, or async runtimes
Bonus
- Experience with Elixir/OTP - GenServers, supervisors, the actor model
- Familiarity with LLM APIs and streaming protocols (SSE, chunked responses)
- You've built or worked on conversational AI systems (not just chatbot wrappers)
- Experience with NATS, Kafka, or other messaging systems
- You've read our whitepaper on Stateful Agent Orchestration and can poke holes in it
How We Work
We care about craft obsessively. Your work gets questioned, pulled apart, and rebuilt - not because we're harsh, but because everyone here holds each other to a standard most places don't bother with. We work out of a hacker house in Vasant Kunj. We strongly encourage everyone to be in office.
If that sounds like the only way you'd want to work - let's talk.
Skills Required
state management , Kafka, Distributed Systems, Elixir, Go, SSE
){kind=logo}
