About Me
Senior Data Engineer with 14+ years on petabyte-scale big data platforms. Open Source Engineer at OneHouse, contributing upstream to Apache Hudi and Apache XTable. Previously Cloudera's Spark and Iceberg backline engineer.
Hello
I am Ranga Reddy, a Senior Data Engineer based in Bangalore, India, with 14+ years of experience building petabyte-scale big data platforms and lakehouse architectures.
Today I work as an Open Source Engineer at OneHouse, contributing upstream to Apache Hudi and Apache XTable: indexing strategies, compaction, clustering, metadata table optimization and multi-writer concurrency. Before that I spent four and a half years at Cloudera as the Spark and Iceberg backline engineer, the final escalation point for the platform’s most critical P0/P1 customer issues across Fortune 500 deployments.
That combination is what this blog is about: format internals on one side, and what actually breaks in production on the other.
What I work on
- Open source development and community work, on Apache Hudi and Apache XTable
- Backline engineering: production escalation and root-cause analysis on Spark and Iceberg
- Spark SQL and DataFrame optimization, including Catalyst query tuning
- Lakehouse architecture with Hudi, Iceberg and Hive Metastore
- Real-time and batch pipeline engineering
- Customer-facing solution engineering and technical writing
Experience
Open Source Engineer, OneHouse
Bangalore, September 2024 to present
Contributing upstream to Apache Hudi and Apache XTable, designing and optimizing core components for large-scale lakehouse systems.
- Upstream development. Authored and reviewed pull requests, JIRAs and RFCs across Hudi internals: indexing strategies, compaction, clustering, metadata table optimization and multi-writer concurrency models.
- Customer escalation engineering. Lead escalation engineer for community and enterprise issues filed through Slack, GitHub and the Hudi mailing list, covering Spark, Trino, Presto and Flink integrations.
- Benchmarking framework. Built automated performance and correctness suites comparing Hudi behaviour across Spark, Presto and Trino, used internally to gate releases.
- Cost-saving tooling. Built end-to-end tooling that parses table metadata, query logs and statistics into actionable metrics, letting customers cut up to 40% of their production costs.
- Technical content. Deep-dive posts and tutorials on Hudi internals for the OSS community.
Stack: Java, Scala, Python, Spark, Trino, Presto, AWS (EMR, S3, Athena, Glue), GCP (Dataproc, GCS, BigQuery).
Staff Software Engineer, Spark and Iceberg Backline, Cloudera
Bangalore, March 2020 to August 2024
The final escalation point for Cloudera’s global support organization: the complex Spark and Iceberg production outages that frontline support and field engineers could not resolve.
- Spark escalations. Root-cause analysis on hundreds of P0/P1 Spark issues across CDP and CDH: shuffle failures, OOM patterns, dynamic allocation regressions, Catalyst optimizer edge cases, and Hive-on-Spark and Spark-on-YARN integration breaks.
- Iceberg production support. Metadata corruption, snapshot isolation conflicts, time-travel query regressions, and partition and schema evolution compatibility across Spark 3.x and Hive Metastore, plus Iceberg’s integration with Trino and Presto.
- Internal tooling. A unified log-parsing tool that ingests Spark event logs, application logs and Iceberg metadata files to surface stage-level bottlenecks and table-level inefficiencies. Adopted by the global support team.
- Cross-platform debugging. Integration issues between Spark, Iceberg and cloud storage (S3, GCS), including consistency models and file-listing performance.
- Knowledge sharing. Published Spark and Iceberg knowledge articles on the Cloudera Community, and designed Spark Hogwarts, a hands-on internal training program on Spark internals and lakehouse formats for the frontline support team.
Stack: Java, Scala, Python, Spark 2.x and 3.x, Kafka, Iceberg, Hive, Cloudera CDP/CDH/HDP.
Big Data Engineer, Dell EMC
Bangalore, November 2015 to March 2020
Support Assist Intelligence Engine (SAIE) and SAIE Analytics: storing Support Assist telemetry to enable proactive and predictive support.
- Designed a Spark to Kafka to Storm to HBase ingest path for unstructured and semi-structured telemetry at scale.
- Read from Oracle, processed through Spark, published to Kafka, consumed from Kafka with Storm for real-time analysis.
- Co-led the HDP 2.3 to 2.5 production upgrade across a 40-node cluster with zero unplanned downtime.
- Led the design of the Support Assist Enterprise module and its configuration ingestion pipeline.
Stack: Hadoop, Hortonworks HDP, Spark, Storm, Kafka, Oracle, HBase.
Senior Software Engineer, Mindtree
Bangalore, August 2014 to November 2015
Meeting Services Application: a mobile and tablet platform letting meeting planners request services without leaving the room. Controller, service and DAO layers, Spring service interfaces, JSP with jQuery and AngularJS, Hibernate.
Stack: Java, Spring, Hibernate, AngularJS, jQuery, MySQL.
Junior Software Engineer, ITApp Software
Bangalore, October 2012 to July 2014
AppCenter: a cloud services delivery and business management platform for provisioning, lifecycle, monitoring and user management. Spring Portlet MVC portlets, controllers, JSP front end, Kendo UI, JUnit.
Stack: Java, Spring, Hibernate, MySQL, Kendo UI, jQuery.
Core skills
| Domain | Technologies |
|---|---|
| Lakehouse and table formats | Apache Hudi (contributor), Apache Iceberg, Hive Metastore |
| Distributed compute | Apache Spark (SQL, Streaming, DataFrames), Apache Storm, Apache Hadoop |
| Streaming and messaging | Apache Kafka, Hudi Streamer |
| Query engines | Trino, PrestoDB, Athena, Spark SQL, Hive |
| Cloud and storage | AWS (S3, Athena, Glue, EMR), GCP, MinIO |
| Languages | Java, Scala, Python |
| Frameworks | Spring, Hibernate |
| Databases | Oracle, MySQL, HBase |
| Distributions | Cloudera CDP/CDH, Hortonworks HDP, Onehouse |
| Tooling | Maven, Gradle, Jenkins, Git, IntelliJ |
Open source
- Apache Hudi: upstream commits across indexing, compaction, clustering, the metadata table and multi-writer concurrency.
- Apache XTable: cross-platform lakehouse interoperability.
- Benchmarking framework: automated performance and correctness validation for Hudi across Spark, Presto and Trino.
- Cloudera Community articles: Spark and Iceberg knowledge base articles.
Education
- Master of Computer Applications (MCA), Sri Krishnadevaraya University, Anantapur, 2011, 80%
- Bachelor of Science (B.Sc), Sri Venkateswara University, Tirupati, 2008, 81%
Recognition
- Silver Award, Dell EMC (Q4 FY18)
- Cash Award for the SAIE Analytics project, Dell EMC
- A-Team Performer for the MSA project, Mindtree
- Cloudera Hackathon and Dell Ideafest participant
Get in touch
- Email: rangareddy.avula@gmail.com
- LinkedIn: linkedin.com/in/rangareddy1988/
- GitHub: github.com/rangareddy
- Stack Overflow: stackoverflow.com/users/1918962/ranga-reddy
Full CV: download my resume (PDF).