Senior Voice AI Platform Engineer
Aivar Innovations Private Limited2 months ago
Bengaluru, IndiaSenior
Responsibilities
- Own distributed, real-time voice AI systems supporting thousands of simultaneous calls.
- Improve throughput and calls-per-vCPU by extending Pipecat or rewriting hot paths in Go.
- Decompose and optimize end-to-end latency across STT, LLM, TTS, RAG, tool-calling, turn detection, and network hops.
- Diagnose pipeline issues and own fixes for calls that do not meet quality or latency expectations.
- Build voice-to-voice evaluations that gate releases and detect provider drift.
- Make STT, LLM, TTS, RAG, and tool-calling provider tradeoffs based on voice quality and latency.
- Extend the platform into live assist, agent assist, and related products.
- Review code, mentor engineers, and define engineering standards for a growing team.
Requirements
- 5+ years building production products and systems, including at least 2 years building and operating real-time voice AI in production.
- Deep understanding of the complete voice pipeline and its effects on quality and latency.
- Strong Go and Python skills, with particular value placed on production-grade, performance-critical Go.
- Experience scaling Pipecat or a comparable real-time voice framework to high concurrency in production.
- Fluency in tradeoffs among STT, LLM, TTS, RAG, and tool-calling providers.
- Experience building or running evaluations for a real-time voice or LLM system using measurable quality signals.
- Fluency with agentic coding tools.
- Applicants whose experience is limited to prompt engineering, LLM API applications, managed voice platforms, CRUD web backends, or ML model training are not the target profile.
Benefits
- Bangalore office location.