Join Sarvam, a pioneering company building India's sovereign AI platform. We are developing a full-stack AI solution encompassing research, models, infrastructure, and applications, with a clear mission to make AI impactful for India. Sarvam collaborates with leading enterprises and public institutions, supported by prominent investors.

Our Model API platform powers developers and enterprises, handling millions of daily API calls across various services like ASR, TTS, Vision, LLM, and Translate. This platform is crucial for shipping products built on our foundation models and includes features like metering, billing, and rate-limiting.

We are seeking a talented Software Engineer to take ownership of key components within this platform, working closely with our Staff Engineer on architectural decisions. This role is ideal for individuals passionate about optimizing system performance, reducing latency, and leveraging AI in their engineering practices.

As a Software Engineer on our Model API Platform team, you will be responsible for the end-to-end development, operation, and enhancement of critical platform areas. This includes owning components such as WebSocket streaming, speech recognition pipelines, or text-to-speech pipelines.

You will transform abstract goals into concrete plans, defining milestones and delivering results while meticulously documenting decisions and trade-offs. Accountability extends post-launch, involving on-call duties, incident response, and root-cause analysis.

A deep focus on latency optimization is paramount. You will dissect existing services, identify performance bottlenecks, and tune system parameters to achieve sub-100ms latency Service Level Objectives (SLOs) in a multi-tenant environment. Building robust real-time and model-serving pipelines, including streaming infrastructure and audio processing, will be a core part of your role.

The ideal candidate possesses 4-6 years of experience in building and operating production backend systems, with a proven track record of owning services end-to-end under significant load. A strong foundation in distributed systems, including concurrency, networking, queues, caching, consistency, and failure modes, is essential.

Experience with low-latency or high-throughput services, and the ability to reason about p95 and p99 latency, is required. Proficiency with relational databases (e.g., PostgreSQL) and caches (e.g., Redis), including schema design and query optimization, is necessary. Fluency in at least one backend language like Python or Golang is expected.

Familiarity with containerization, CI/CD pipelines, and production monitoring is important. We value engineers who actively use AI tools in their workflow to automate or accelerate tasks and demonstrate a bias for action with clear, concise written communication, particularly for RFCs and design documents.