Member of Technical Staff (AI Supportability)

Cockroach Labs
Bangalore, IndiaPosted 26 February 2026

Job Description

<div class="job-post-container" style="position: relative; background-color: white; padding: 20px; max-width: 1000px; margin: 0 auto; box-shadow: 0px 0px 10px rgba(0, 0, 0, 0.1); min-height: 500px; border: 5px solid #0037A5;"> <div style="position: absolute; top: 0; left: 2; width: 200px; height: 300px; margin-top: 30px; opacity: 1; z-index: 1;"> </div> <div class="job-post-content" style="position: relative; z-index: 2;"> <h1 style="font-family: Arial, sans-serif; color: rgb(51, 51, 51); line-height: 1.2; margin-top: 20px; text-align: center;"><img style="max-width: 100%; float: left;" src="https://i.imgur.com/b9m4CJj.jpeg" alt="" width="92"></h1> <h1 style="font-family: Arial, sans-serif; color: rgb(51, 51, 51); line-height: 1.2; margin-top: 20px; text-align: center;"> </h1> <h1 style="font-family: Arial, sans-serif; color: rgb(51, 51, 51); line-height: 1.2; margin-top: 20px; text-align: center;"> </h1> <h1 style="font-family: Arial, sans-serif; color: rgb(51, 51, 51); line-height: 1.2; margin-top: 20px; text-align: center;">Category-defining tech. Career-defining work.</h1> <p style="text-align: center;">Lots of tech companies disrupt. But, many fail when they try to scale. We're different. CockroachDB makes it easier for companies to build and scale apps. This is how and why we're helping some of the most innovative companies on the planet. We tackle problems head-on and focus on solutions that create lasting impact. </p> <h3 style="text-align: center;">Because when our customers win, we all win. </h3> <p> </p> <hr> <h2><span style="color: rgb(0, 0, 0);"><strong>The Role</strong></span></h2> <p>Can LLMs accurately diagnose complex database issues from telemetry data? We're building a team to prove this.</p> <p>Supporting CockroachDB is inherently challenging. Our product has thousands of unique metrics, complex structured data, and emits detailed logs at high frequency, making support tickets expensive and time-consuming to resolve. We're launching an RD initiative to explore how Applied AI can reduce these support costs and improve diagnostic accuracy.</p> <p>We're looking for engineers who thrive in exploratory, prototype-driven environments working closely with end users. You'll start with a concrete mission: validate whether LLMs can perform root cause analysis on real Zendesk escalations using CockroachDB telemetry data. Working embedded with our technical support engineers, you'll experiment with OpenAI/Anthropic APIs, prompt engineering techniques, and various data contexts to determine if AI can transform diagnostic workflows.</p> <p>This isn't about executing a predetermined roadmap. It's about rapid experimentation, learning from real data, and building the foundation for AI-powered support workflows. If you're energized by building from scratch, working directly with internal stakeholders, and technical challenges where the outcome isn't guaranteed, this role is for you.</p> <h2><strong>You Will</strong></h2> <ul> <li><strong>Start with focused validation</strong>: Design experiments to test whether LLMs can assist with root cause analysis on real Zendesk escalations, determining if AI can accelerate diagnostic workflows and reduce time-to-resolution.</li> <li><strong>Work embedded with our Bangalore technical support team: </strong>Partner with support engineers through regular collaboration sessions to understand diagnostic workflows, review experimental results, and gather insights.</li> <li><strong>Work with real CockroachDB data</strong>: Process diagnostic telemetry including cluster metrics, SQL traces, system logs, plus observability data.</li> <li><strong>Iterate on AI approaches</strong>: Experiment with OpenAI/Anthropic APIs, prompt engineering techniques, and different data contexts to optimize LLM diagnostic performance.</li> <li><strong>Build rapid prototypes</strong>: Transform promising validation results into proof-of-concept tools and agents, defining next steps based on what the data tells us. ... (truncated, view full listing at source)