Senior Platform Engineer

Walt is an AI data-analyst platform. Enterprise customers connect their data warehouses — BigQuery, Snowflake, Databricks, Trino, Postgres — and Walt's agentic engine answers business questions, builds semantic models over their schemas, and generates analyses and presentations on top of live data. Because our customers' data is highly sensitive, we run a split deployment model: a multi-tenant SaaS on our own Kubernetes clusters, plus dedicated deployments inside customers' own cloud accounts (Azure and GCP today), orchestrated from our central GitOps control plane. You will be the owner of that machinery.

Responsibilities:

• Own the deployment engine: Our fleet runs on GKE across dev, staging/demo, and prod clusters, deployed via Helm charts and ArgoCD with per-customer namespaces and environments. You'll harden, simplify, and scale this GitOps pipeline.
• Build the customer-cloud (BYOC) path: Productize how Walt ships into a customer's own cloud — today hand-crafted Helm charts, cloud-native secret stores (Azure Key Vault CSI, GCP Secret Manager), and customer-side CI. Turn it into a repeatable, self-serve delivery engine.
• Master cross-account security: Manage per-customer encryption keys for connector credentials, GCP service accounts and Azure service principals across org boundaries, workload identity, and network isolation between customer environments.
• Harden network data paths: Implement and maintain private, secure networking paths using GCP Private Service Connect (PSC) and Azure Private Link to connect securely to customer warehouses without traversing the public internet.
• Run the production stack: Postgres, Redis, Celery workers, and sandboxed compute (Daytona) under real workloads — capacity planning, autoscaling, OOM and incident response, database migration safety, preview environments for every PR.
• Enable enterprise sales & trust: Produce the security architecture docs and isolation blueprints that get Walt through CISO and procurement reviews. Act as the technical authority on architecture calls with early enterprise IT teams to unblock deployments.

Requirements:

• Kubernetes depth: GKE (or EKS/AKS) in production — autoscaling, node-pool and resource management, StatefulSets, ingress/TLS, multi-cluster operations.
• GitOps and CI/CD fluency: ArgoCD (or Flux) and GitHub Actions as the backbone of deployment; Helm chart authoring beyond the basics.
• Cross-cloud IAM expertise: GCP service accounts and Azure service principals/Key Vault across organisation boundaries; secrets lifecycle and key rotation.
• IaC and Automation skills: Terraform/OpenTofu for programmatically standing up customer environments. Familiarity with the Kubernetes Operator pattern or CRDs to automate self-serve infrastructure delivery is a massive plus.
• Infrastructure Coding / Scripting: Strong coding capability in Python or Go to build custom orchestration tooling, optimise Celery worker queues, and safely debug backend application issues.
• Data-platform sympathy: Comfort operating systems that connect to BigQuery, Snowflake, Databricks, and friends — credentials handling, network paths, query workloads.
• Startup mindset: You ship stable infrastructure over theoretical perfection and are comfortable as the sole infra owner working directly with founders.

Bonus Points:

• Built a BYOC or managed single-tenant delivery model before.
• Cloud marketplace listings (GCP/Azure/AWS).
• Experience with AI/LLM workloads: sandboxed code execution, token/cost controls, eval pipelines.
• Postgres operations at depth (migrations, backup/restore, encryption at rest).

Skills

Argo CD, Helm, Platform Engineer