Database Architect

Title: Data Platform / Database Architect (Postgres + Kafka) — AI‑Ready Data InfrastructureLocation: Noida (Hybrid). Remote within IST±3 considered for exceptional candidates.Employment: Full‑timeAbout UsWe are building a high‑throughput, audit‑friendly data platform that powers a SaaS for financial data automation and reconciliation. The stack blends OLTP (Postgres), streaming (Kafka/Debezium), and OLAP (ClickHouse/Snowflake/BigQuery), with hooks for AI use‑cases (vector search, feature store, RAG).Role SummaryOwn the end‑to‑end design and performance of our data platform—from multi‑tenant Postgres schemas to CDC pipelines and analytics stores—while laying the groundwork for AI‑powered product features.What You’ll Do• Design multi‑tenant Postgres schemas (partitioning, indexing, normalization, RLS), and define retention/archival strategies.  • Make Postgres fast and reliable: EXPLAIN/ANALYZE, connection pooling, vacuum/bloat control, query/index tuning, replication.  • Build event‑streaming/CDC with Kafka/Debezium (topics, partitions, schema registry), and deliver data to ClickHouse/Snowflake/BigQuery.  • Model analytics layers (star/snowflake), orchestrate jobs (Airflow/Dagster), and implement dbt‑based transformations.  • Establish observability and SLOs for data: query/queue metrics, tracing, alerting, capacity planning.  • Implement data security: encryption, masking, tokenization of PII, IAM boundaries; contribute to PCI‑like audit posture.  • Integrate AI plumbing: vector embeddings (pgvector/Milvus), basic feature‑store patterns (Feast), retrieval pipelines and metadata lineage.  • Collaborate with backend/ML/product to review designs, coach engineers, write docs/runbooks, and lead migrations.Must‑Have Qualifications• 6+ years building high‑scale data platforms with deep PostgreSQL experience (partitioning, advanced indexing, query planning, replication/HA).  • Hands‑on with Kafka (or equivalent) and Debezium/CDC patterns; schema registry (Avro/Protobuf) and exactly‑once/at‑least‑once tradeoffs.  • One or more analytics engines at scale: ClickHouse, Snowflake, or BigQuery, plus strong SQL.  • Python for data tooling (pydantic, SQLAlchemy, or similar); orchestration with Airflow or Dagster; transformations with dbt.  • Solid cloud experience (AWS/GCP/Azure)—networking, security groups/IAM, secrets management, cost controls.  • Pragmatic performance engineering mindset; excellent communication and documentation.Nice‑to‑Have• Vector/semantic search (pgvector/Milvus/Pinecone), feature store (Feast), or RAG data pipelines.  • Experience in fintech‑style domains (reconciliation, ledgers, payments) and SOX/PCI‑like controls.  • Infra‑as‑Code (Terraform), containerized services (Docker/K8s), and observability stacks (Prometheus/Grafana/OpenTelemetry).  • Exposure to Go/Java for stream processors/consumers.  • Lakehouse formats (Delta/Iceberg/Hudi).