Staff Data Engineer

About the role

Sarvam is building a sovereign AI platform for India and needs a Staff Data Engineer to design and build its Data & Analytics Platform from the ground up. This high‑ownership role will be the technical owner of a system that every Sarvam product writes to and every customer reads from, handling event ingestion, CDC replication, and tenant‑aware analytics.

Key responsibilities

• Design and implement a high‑throughput event ingestion API in Go that validates arbitrary payloads and stores them in ClickHouse with automatic schema evolution.
• Build CDC pipelines using Debezium and Kafka to mirror PostgreSQL/MySQL product databases into ClickHouse, ensuring deduplication, ordering, and replay tooling.
• Develop a multi‑tenant query API with strict RBAC, per‑tenant isolation, and query budgeting, powering dashboards and finance billing extracts.
• Create a no‑code dashboard layer that enables customers and internal teams to build analytics views without writing SQL.
• Architect and operate the ClickHouse cluster (sharding, replication, MergeTree families, materialized views) with capacity planning, cost controls, observability, and SLOs.
• Define engineering culture for the team, including design reviews, RFC process, testing standards, and hiring criteria.

Required profile

• 5+ years of backend or data‑platform engineering experience, with a proven track record of building 0‑to‑1 platforms used by multiple engineering teams.
• Hands‑on experience operating ClickHouse at scale, including sharding, replication, MergeTree variants, and query tuning.
• Strong proficiency in Go for high‑throughput services and API design (REST/gRPC).
• Deep knowledge of Kafka as a streaming backbone, including partitioning, consumer groups, and exactly‑once semantics.
• Experience building Debezium‑based CDC pipelines from transactional databases to analytical stores.
• Solid systems instincts for throughput, latency, cost, and failure‑mode analysis.

Required skills

• ClickHouse
• Go / Golang
• Kafka
• Debezium
• PostgreSQL
• MySQL
• Spark (optional)
• Airflow (optional)
• Superset, Metabase, Cube (optional)