Staff Data Engineer

Sarvam is building an Indian sovereign AI platform used by enterprises including Tata Capital, SBI Life and CRED, and this role builds the data and analytics platform underneath all of it from the ground up. The posting is unusually concrete about the mandate. The platform has three jobs: ingest every meaningful event from every Sarvam product, covering outbound calls, agent turns, user turns, knowledge base lookups, lead creation, deployments and model invocations, through an API layer that handles automatic schema evolution and dynamic table creation per event type; replicate product databases into the analytics store using change data capture so operational data stays queryable without loading the transactional systems; and expose all of it through a tenant aware query API and a no code dashboard layer that powers customer facing analytics, internal reporting and the finance team's billing pipeline. A firm architectural rule runs through it: databases, Kafka and internal stores are never exposed directly, and every read and write goes through APIs you design. You would own the architecture, build the core hands on, and shape the team that grows around it.

What the posting requires

- 5+ years of backend or data platform engineering, with a meaningful part of it designing and operating analytics systems in production
- A track record of building zero to one platforms that other engineering teams depend on
- ClickHouse in production at scale: sharding, replication, MergeTree variants, materialized views, projection design, query tuning, and the operational side including backups, upgrades and capacity planning
- Go for high throughput backend services, with a strong grasp of concurrency, performance and API design across REST and gRPC
- Kafka as a primary streaming backbone: partitioning strategy, consumer groups, exactly once and idempotency patterns, and schema registry
- Debezium based change data capture, built and operated from transactional databases into analytical stores
- Strong systems instincts, reasoning about throughput, latency, cost and failure modes before writing code

Counts as a bonus, not a requirement

- Spark and Airflow for batch enrichment, backfills and scheduled aggregation
- Multi tenant SaaS platforms with row level security, per tenant quotas and tenant aware access control
- Building or integrating BI and dashboard tools such as Superset, Metabase or Cube, especially embedded customer facing analytics
- Exposure to billing or metering data feeds where correctness is not negotiable

What you would actually be building

- A high throughput event ingestion API in Go that accepts arbitrary event payloads, validates them, and lands them in ClickHouse with automatic table creation
- Change data capture replication from product databases into the analytics store
- A tenant aware query API and no code dashboard layer feeding customer analytics, internal reporting and billing

Location and working pattern

On site in Bengaluru, working with a small team of senior engineers. The posting says this directly, so treat it as an on site role rather than hybrid.

A good fit if

You want to be the technical owner of a foundational platform rather than a feature contributor on someone else's system, you like writing design docs as much as code, and you think about cost per query as well as latency. Staff here means senior individual contributor, so the scope is architectural.

Think twice if

You prefer a defined slice of an existing system, or you need hybrid or remote flexibility.