Software Engineer - Model API Platform

Own major components of Sarvam's model API platform end to end, such as speech-to-text or text-to-speech pipelines, ensuring reliability, performance, and low latencies.

Build and operate real-time streaming infrastructure including WebSocket layers and audio processing pipelines targeting sub-100ms latency SLOs.

Lead technical decisions, incident response, root cause analysis, and continuous improvement while collaborating with MLOps and Inference teams.

4-6 years of experience building and operating production backend systems with at least one end-to-end owned service under real load.

Strong fundamentals in distributed systems: concurrency, networking, queues, caching, consistency, and failure modes.

Proficiency in backend development with fluency in at least one language; experience in Python and Golang preferred.

Experience with relational databases (e.g., PostgreSQL), caching (e.g., Redis), containerized services, CI/CD, and production monitoring.

Experienced in low-latency, high-throughput backend services and comfortable analyzing p95/p99 latency metrics.

Skilled at system thinking from first principles including latency budgets, throughput, failure modes, and cost trade-offs.

Has operational ownership mindset: writes RFCs and design docs, mentors juniors, and integrates observability and testing for fast, reliable shipping.