Presented by:

Hari Krishna Sunder

YugabyteDB

Hari Krishna Sunder is the Technical Lead and Principal Architect at YugabyteDB, specializing in cloud-native database systems and distributed architectures.

With over 13 years of experience, Hari has built core database components for major systems including YugabyteDB, Azure SQL Database, and Azure Synapse Data Warehouse. He currently serves on the Technical Steering Committee for DocumentDB, a CNCF project backed by Microsoft, YugabyteDB, and Amazon.

An active member of the open-source community and various database conference program committees, Hari’s expertise lies in database internals such as storage, Write-Ahead Logging, replication, recovery, and metadata management, with a focus on delivering performance and reliability at scale.

No video of the event yet, sorry!

DocumentDB ** DocumentDB **

Developers are increasingly tasked with building cloud ready applications that require both flexible data schemas and advanced AI capabilities. Traditionally, this meant managing a complex stack of document stores, vector databases, and relational engines. PostgreSQL has been the de facto solution for the latter two, and now with DocumentDB it has the power to take on the entire stack.

DocumentDB is built on the principles of transparency, developer freedom, and standardization, with a mission to build a MongoDB compatible open source document database based on PostgreSQL. Donated to the Linux Foundation by Microsoft, the project is backed by its creator alongside industry leaders Amazon, YugabyteDB and many others.

Key technical topics include

  • Custom Data Types: Implementing a native BSON data type for efficient storage and retrieval.
  • The API Layer: How the pg_documentdb_api translates and runs CRUD operations and advanced queries in PostgreSQL.
  • Vector Integration: How we integrated pg_vector to enable semantic search over unstructured document data.
  • Indexing at Depth: Techniques for implementing field-level indexing at any nesting depth using PostgreSQL’s GIN and GiST indexes.

This session is designed for developers and architects interested in the internal mechanics of PostgreSQL extensions and those looking to maximize their development workflow without leaving the PostgreSQL ecosystem.

Date:
2026 April 23 14:30 PDT
Duration:
50 min
Room:
Winchester
Conference:
Postgres Conference: 2026
Language:
Track:
Postgres Extensions Day
Difficulty:
Easy