Platform Engineer (Terraform|K8s|Data Infra)

Company Description

SentiSum is an AI-native Voice of the Customer platform that acts as a Customer Second Brain, consolidating feedback across interactions like support tickets, chat logs, surveys, and reviews. Its real-time AI identifies root causes, patterns, and anomalies to provide actionable insights and prevent problems before they escalate.

The role

• You'll own the infrastructure our AI runs on - and you'll own it end to end.

• You'll be the person who decides how we run things, and it starts with Terraform: infrastructure as code for everything.

• Kubernetes to run and scale the workloads, and CI/CD that makes shipping routine rather than risky.

The first big job on that foundation is getting our data orchestration pipelines solid: the workflows that pull customer feedback in, push it through our AI, and surface insights on the other side. They need to be reliable, observable, and pleasant for the rest of the team to build on.

The role expands as our AI does. Increasingly that means LLMOps - the infrastructure that runs LLMs and AI agents in production, and operating it well: cost, rate limits, scale, and observability. It's a real growth path if you want to move toward AI infrastructure.

You'll probably love this if..

• You manage infrastructure as code with Terraform.
• Kubernetes in production doesn't scare you - you've debugged the 2am pod, and lived to automate it away
• You've run a real orchestrator (Airflow, Dagster, or similar) and have opinions about DAGs, retries, and backfills.
• You treat observability as part of the build, not an afterthought - metrics, traces, and alerts that catch problems before users do (OpenTelemetry, SigNoz, Langfuse, or similar)
• You think the best ops work is the work nobody notices, because nothing broke
• You've built the kind of CI/CD and developer tooling that makes a whole team faster
• I'll just automate it is a reflex, not a stretch goal

Bonus points

• You've wrangled data pipelines at scale and know where they tend to rot
• You've touched LLMOps / AI infra - model serving, agent infrastructure, GPU workloads - or you're hungry to
• You've run production on Azure (where we live), though strong K8s/Terraform skills travel fine