Founding ML Infrastructure Engineer

About the Role

We're hiring a Founding ML Infrastructure Engineer in Bengaluru to own the systems that train and serve our Speech models: the distributed training stack, Realtime inference engine, and cloud services. You'll work in a small team with abundant compute resources.

What You'll Do

• Own the training stack: Implement FSDP/HSDP and Megatron-style tensor parallelism within our PyTorch framework.
  
• Make realtime inference fast: Optimize time-to-first-audio and throughput on vLLM-based engines, handling quantization, model compilation, and load testing under high concurrency.

• Build data pipelines at scale:Create in-house pipelines processing millions of hours of audio for high-throughput training feeds.

• Run the production path: Develop API gateways, streaming WebSocket APIs, deploy tooling, logging, and monitoring solutions.

We're Looking For

• 3+ years building ML systems or high-performance backend infrastructure.
  
• Hands-on depth in distributed training (FSDP/TP/PP), LLM serving (vLLM/SGLang/TensorRT-LLM), High volume data processing, or low-latency streaming systems.

• Comfort operating in ambiguity: ability to spec, build, debug, and ship.
  
• Note: Speech/audio experience is a plus but not required.