Overview

Building India's most powerful custom SLMs and real-time Speech-to-Speech voice agents that already handle millions of calls across banking, healthcare, and new economy brands.

What you'll build

Architect infrastructure for custom SLMs and real-time Speech-to-Speech voice agents
Build and scale custom diffusion-based audio models
Optimize inference to handle 10,000+ requests per second
Deploy models using TorchScript, ONNX, TensorRT, and DeepSpeed

Who you are

Experience building or contributing to diffusion-based audio models
Deep expertise in PyTorch
Production deployment experience with TorchScript, ONNX, or TensorRT
Experience optimizing inference at massive scale (10,000+ RPS)

Why this

Models that touch real people — banking, healthcare, new economy brands at millions of calls
Ownership over audio infrastructure at a scale most engineers never reach