Join Sarvam, a leader in building India's sovereign AI platform, dedicated to making AI truly impactful for the nation. We collaborate with premier enterprises and institutions, backed by top-tier investors. Our model APIs are the gateway for developers and businesses to leverage our foundational AI models, handling millions of daily requests.

This role offers a unique opportunity to own the end-to-end development, reliability, and performance of our critical API platform. You will be instrumental in shaping its architecture and ensuring its evolution.

Spearhead the comprehensive design and ongoing development of the API platform, managing the entire request lifecycle. Lead the crucial Python to Go rewrite initiative, defining architectural standards and ensuring seamless migration without compromising reliability. Engineer sophisticated audio pipelines, including TTS chunking, sample-rate adjustments, format encoding, and VAD-based silence detection for ASR. Develop robust vision pipelines for structured data extraction, orchestrating OCR, layout detection, and VLM functionalities. Build and optimize streaming infrastructure, focusing on WebSocket connections, queue-based batch processing, back pressure, and low-latency model invocation. Enhance the commercial layer, ensuring metering, billing, prepaid wallet management, and rate-limiting meet the highest standards. Drive performance and reliability for a multi-tenant platform, consistently achieving sub-second latency Service Level Objectives (SLOs). Implement comprehensive observability solutions (logging, metrics, tracing) to empower confident and rapid production debugging. Develop integration testing infrastructure to facilitate rapid, safe releases of customer-facing APIs. Foster close collaboration with Inference and MLOps teams, mentoring them on production engineering best practices for fast, reliable, and operable model serving. Champion a design-first, documentation-first engineering culture, emphasizing RFCs, documented decisions, and knowledge sharing.

We are seeking an experienced Staff Engineer with a minimum of 7 years of experience in building scalable production backend systems. Proficiency in Go is essential, with a proven track record of designing, building, and operating Go services under significant load. Familiarity with Python is required, as you will be involved in both languages during the ongoing rewrite, with FastAPI serving as the current foundation. Demonstrated ability to design and operate high-scale, low-latency, multi-tenant distributed systems is crucial. Hands-on experience with real-time and streaming systems, including WebSockets, long-lived connections, and backpressure management, is expected. Strong fundamental understanding of PostgreSQL and Redis, encompassing schema design, query optimization, and caching strategies. Comfort and experience running services on Kubernetes in a production environment. Exhibit strong judgment as a senior individual contributor, discerning when to build versus buy, when to optimize versus ship, and when to abstract versus inline, while actively mentoring others.