Virag Tripathi

Hi, I’m Virag.

Sales engineering. Applied AI. Hands-on delivery.

From customer
challenge to
production.

I work with customers and partners to turn complex problems into working AI and data solutions—from discovery and proofs of concept to deployment and adoption.

Principal, AI Ecosystems Engineer · Cockroach Labs

Explore a connection. See the work behind it.

How I work

Start with the customer.
Build the way forward.

My work sits across sales engineering, solution architecture, and applied AI. I stay close to the problem and hands-on with the solution.

01 / Discover

Understand what matters.

Work with customers and partners to understand the use case, the existing architecture, and what a successful evaluation needs to demonstrate.

02 / Prove

Make it tangible.

Build a proof of concept, integration, or technical demo. Test the difficult parts with the people who will use it.

03 / Deliver

Carry it into production.

Work through deployment blockers with engineering, turn field feedback into product improvements, and help teams adopt the solution.

01 / Selected work

Built to be used.

Three stories about turning a missing connection into something people can build on.

Change data capture / two directions
CockroachDB Debezium
source
Kafka
Kafka JDBC
sink
CockroachDB
Source connector + sink support
Published on Confluent Hub
01 / Data integration

A database connector.
An upstream contribution.

I built the CockroachDB source connector and JDBC sink support, giving data teams a familiar way to move changes into and out of CockroachDB.

Adopted upstream. Source and sink published on Confluent Hub.

Read the story
Developer tools / meet your database
Explore the schema.
Explain this query.
Build the next feature.
Codex Claude Code Cursor Copilot
02 / AI developer experience

Your database.
Inside your coding assistant.

I built CockroachDB plugins for AI coding assistants, bringing database guidance and connected tools into the place developers already work.

Plugins, shared skills, and a published ChatGPT integration.

Read the story
Platform engineering / across architectures
Z LinuxONE · s390x
Power IBM · ppc64le
Validation → platform support → certification
03 / Platform & partner engineering

From platform validation
to an IBM partnership.

I led the technical work to bring CockroachDB onto IBM LinuxONE and Power, connecting platform engineering with a partnership that reached customers.

Technical groundwork supporting IBM’s 2025 OEM agreement.

Read the story

02 / See it work

Plain English.
Real database work.

A short walkthrough of the CockroachDB plugin in Codex, using synthetic demo data.

  1. 01Explore the schema
  2. 02Query data and inspect the plan
  3. 03Create, insert, and verify with UUID keys
Read the launch story ↗
68 seconds · Silent walkthrough · Edited from my Codex session
What happens in the walkthrough

I ask Codex to explore the demo database and its table schema, read the five newest synthetic samples, and explain the query plan. I then ask it to create a scratch table with generated UUID primary keys, insert two synthetic notes, and read them back. The recording shows the tool responses and resulting data. Waiting time and a recovery detour are cut from this short edit.

03 / More from the workbench

Keep exploring.

Integrations, experiments, and tools I've built or contributed to.

AI

LangChain + LangGraph

Vector and hybrid search, chat history, checkpoints, and long-term agent memory. I designed, wrote, and maintain the integration.

AI

Google MCP Toolbox

I wrote the CockroachDB integration and worked with Google to merge it upstream.

Data

CockroachDB → ClickHouse

A native PeerDB connector for CDC into ClickHouse. Validated with node failures during a full mirror.

Data

Apache Airflow compatibility

A compatibility patch I maintain for running standard Airflow on CockroachDB, including scheduler retry handling.

Data

VectorDBBench

I added CockroachDB support to the open-source vector database benchmark.

Tools

Tools for the work between

Schema and data exports, JDBC retries, and database diagnostics: crdb-dump, a JDBC wrapper, and an analyzer.

04 / A little context

Customer-facing.
Hands-on.

I've been a Solutions Architect, a Principal Field Engineer, and now an AI Ecosystems Engineer. Sales engineering is a core part of that story: understanding a customer's problem, earning technical confidence, and helping a team get a solution into production.

At Redis, I paired that work with commercial results: 160% of ARR goal in 2020 and the most net-new logos in the Americas. At Cockroach Labs, I've led technical work across IBM, core banking, CDC, and AI partnerships.

I still write the code myself. It helps me test the idea, work through the blockers, and build something real with the customer.

Read my resume →
  1. 2026–present
    Cockroach LabsPrincipal, AI Ecosystems Engineer
  2. 2024–2025
    Cockroach LabsPrincipal Solutions Architect
  3. 2018–2024
    RedisSolutions Architect → Principal Field Engineer
  4. 2000–2018
    Before thatTIBCO · JetBlue · Delta · Cingular

IBM Champion · Cockroach Labs President's Club · Redis CEO Award and President's Club

05 / Ideas in the open

Writing & speaking.

What I learned building it, and how you can use it.

Let's build something useful

The best conversations
start with a real problem.

Get in touch