Data Architect — Databricks

Data Engineering & Pipelines | Mid-Level | Full-Time

Experience

5 – 8 Years

Level

Mid-Level

Employment
Type

Full time

Location

Pune - Hybrid

Primary
Stack

Databricks,
Apache Spark, Delta Lake, SQL

Domain

Data
Engineering & Pipelines

About the Role

We are looking for a hands-on Data
Architect with deep expertise in Databricks to design, build, and optimise
enterprise-scale data platforms. You will own the end-to-end data engineering
lifecycle — from ingestion and transformation to serving — while ensuring
reliability, scalability, and governance across our lakehouse architecture.

You will collaborate closely with
data engineers, analytics engineers, and product teams to translate business
requirements into robust, reusable data solutions on the Databricks Lakehouse
Platform.

Key Responsibilities

Data
Architecture & Design

• Design and maintain the
organisation's lakehouse architecture using Databricks and Delta Lake.

• Define data modelling
standards (dimensional, Data Vault 2.0, or medallion architecture) across
Bronze,



Silver, and Gold layers.

• Architect scalable
ingestion frameworks using structured and unstructured data sources (Kafka,
JDBC, REST APIs, cloud storage).

• Own schema evolution
strategy and ensure backward-compatibility across data assets.

Pipeline
Development & Optimisation

• Build and maintain
production-grade ETL/ELT pipelines using PySpark, Spark SQL, and Databricks
Workflows.

• Implement Delta Live Tables
(DLT) for declarative, auto-scaling pipeline development.

• Optimise Spark jobs for
performance — partitioning, Z-ordering, caching, and cluster right-sizing.

• Establish CI/CD practices
for data pipelines using tools such as GitHub Actions, Azure DevOps, or
Databricks Asset Bundles.

Data
Governance & Quality

• Implement Unity Catalog for
data discovery, lineage tracking, fine-grained access control, and compliance.

• Define and enforce data
quality rules using Great Expectations, DLT e

📌 Data Architect (Pune)
🏢 nCircle Tech
📍 Pune