Senior Software Engineer - Reliability

Overview

Join Atlan's Reliability team to build and improve a production multi-agent AI SRE platform. The role focuses on making agent investigations accurate and trustworthy, expanding safe auto-remediation, and enabling other teams to use the platform in production.

What you'll do

• Improve investigation agents' root-cause accuracy so engineers can trust their answers without re-checking.
• Expand auto-remediation from a handful of playbooks to dozens, using staged dry-run, human-approval and autonomous execution.
• Build and run fault-injection benchmarks and evaluation harnesses for agent trustworthiness.
• Set evaluation pass marks, report results with honest denominators, and validate the evaluation harness itself.
• Design fail-closed checks, kill switches, approval flows and blast-radius limits for agents writing to production.
• Build new agents that remove categories of operational toil.
• Build clean interfaces, guardrails and safe defaults so other teams can use the reliability platform in their own production environments.
• Respond to production issues by stopping the bleeding first and then removing the underlying class of failure.

What you'll need

• Experience carrying a pager, owning incidents end to end, or working in a support or escalation queue long enough to understand which operational pain is worth removing.
• Experience eliminating a recurring operational problem rather than only making a runbook faster.
• Ability to identify who used what you built, what changed, and the relevant numbers and denominators.
• Experience rebuilding a core part of your own work end to end with AI so that the workflow is structurally different, not just faster.
• Experience shipping an AI-native workflow that other people depend on.
• Ability to describe guardrails before capabilities when architecting agents that take action autonomously.
• Ability to reason about false-positive rates, rollback paths and blast radius when agents are wrong in production.
• Ability to build clean interfaces, document failure modes and provide safe defaults for someone else's production.
• Ability to contribute quickly to a high-rigor existing codebase and improve it rather than rewrite it.
• Ability to use deterministic code for routing, filtering and safety before generative steps.
• Ability to think in systems and use agents as leverage rather than as a replacement for judgment.
• Genuine interest in reliability work and operational efficiency.

Nice to have

• Experience getting other teams to adopt a platform you built.

Details

• Location: India.
• Work best with overlap between roughly 11am and 8pm IST.