Site Reliability Engineer (Bengaluru)

We Are Hiring | Product Manager, Enterprise AI Platform | Bengaluru

About ReN3

ReN3 is a sovereign AI platform for the enterprise, built to unify an organization's knowledge and let teams execute AI-powered work under strict governance and access control. The platform runs in air-gapped, private-cloud, or other sovereign environments, so sensitive data never leaves the client's own infrastructure. ReN3 works with government, banking and financial services, insurance, healthcare, manufacturing, and construction clients, and partners with major infrastructure and cloud providers on deployment.

About the Role

As a Site Reliability Engineer, you will own the reliability, security and cost of a multi-tenant AI platform running on Kubernetes. You’ll own the full spectrum of reliability engineering, from provisioning infrastructure with Terraform to building alerting systems that keep our services running smoothly. As one of the first SREs on the team, you’ll have the opportunity to shape our engineering practices, establish standards, and build a strong foundation for reliability as we scale.

Our stack

• Platform: GKE, Terraform, Helm charts deployed through GitOps (Argo CD)
• Services: TypeScript (NestJS) APIs and workers, Python agent runtime, JVM data-processing jobs
• Data and messaging: PostgreSQL (plus MySQL and SQL Server for some customer deployments), Kafka, Temporal, Redis, a vector store
• Security: OpenFGA for fine-grained permissions, a secrets manager, sandboxed code execution (gVisor)
• Observability: OpenTelemetry traces, metrics and logs
• AI: several LLM providers behind a gateway, plus self-hosted models

Key Responsibilities

• Own SLOs. Define and defend uptime and latency targets for the APIs, the document-ingestion pipeline and the agent runtime.
• Change infrastructure only through code. Every change goes through Terraform and GitOps: planned, reviewed and applied. No click-ops and no ad-hoc kubectl in production.
• Make the system observable. Build tracing, dashboards and alerts that point to the cause, not just the symptom. Add AI-specific signals: token spend, provider latency and errors, queue depth and workflow backlog.
• Lead incidents. Run on-call, lead response, write blameless postmortems and make sure the root cause gets fixed.
• Control capacity and cost. Autoscaling, Kafka and Temporal throughput, database sizing, and LLM and GPU spend.
• Keep tenants isolated and secure. Secrets handling, network policy, sandbox isolation and least-privilege access. Help us pass customer security reviews and compliance audits.
• Support customer deployments. Some customers need their own region, their own cloud or their own data centre. Help make those installs repeatable.
• Run operations with agents, safely. Use coding agents to write Terraform, runbooks, scripts and alert triage. Build the guardrails (plan review, dry runs, scoped credentials, human approval) that keep agents from making unsafe changes to production.

What We're Looking For

Must-have

• Based in Bengaluru
• 3–5 years in SRE, DevOps or platform engineering, running production Kubernetes for real users.
• Strong Terraform and Kubernetes skills on at least one major cloud (GCP preferred; AWS or Azure is fine).
• You write real code in Python, Go or TypeScript: tools and automation, not only YAML.
• Daily use of agentic coding, including on infrastructure. You can explain how you use agents without risking production.
• Hands-on incident experience with databases, queues, networking or DNS. You have been the one who found the cause.
• Observability experience with OpenTelemetry, Prometheus/Grafana, Datadog or similar.
• High agency. Take initiative, solve problems, deliver results, adapt quickly, and own your outcomes.

Nice-to-have

• Running Kafka or Temporal in production.
• PostgreSQL operations: tuning, backups and point-in-time recovery, upgrades.
• Multi-tenant SaaS for enterprise or regulated customers.
• OpenShift, on-premises or air-gapped deployments.
• Serving LLMs or running GPU workloads.
• Sandboxing technologies such as gVisor or Firecracker.

How to apply

• Submit your resume (more specifications below).
• Ensure your resume includes a link to something you built mostly with coding agents (a repo, a live URL, a writeup or a 3-minute Loom).
• Ensure your resume includes a short paragraph on a production incident you led: what broke, how you found it and what you changed afterwards.