Platform Engineer

What You'll Own

You'll own the infrastructure foundation of Kepler's AI research platform. The cloud, databases, deployment fabric, and security posture that financial institutions stake their workflows on. Every model invocation, every analyst query, every enterprise deployment runs on what you build.

This is the first dedicated platform hire. You're defining how Kepler ships, scales, and earns trust from the most security-conscious buyers in the world. This isn't a service role. It's the role that defines how Kepler ships.

Responsibilities

• Own the AWS account architecture, Pulumi baseline, and deployment pipelines the whole company depends on
• Stand up the enterprise deployment model (single-tenant, in-VPC, BYOK)
• Set the security baseline (SOC 2 controls, audit logging, identity, network isolation) that converts procurement reviews into closed deals
• Establish the paved path: service templates, CI/CD, secrets management, and observability defaults so every engineer ships safely without thinking about it
• Drive Kepler's database strategy at scale: Postgres performance, replication, tenant isolation, schema evolution as we ingest billions of provenance-tagged records
• Own production reliability: SLOs, on-call rotations, incident response, postmortems
• Build systems that earn trust under real load
• Set the architectural patterns that compound as the team grows: capacity planning, cost discipline, and service standards that others reach for

Requirements

• 5+ years building and operating production cloud platforms
• Deep AWS fluency (VPCs, IAM, KMS, EKS or ECS, RDS) and IaC discipline
• Pulumi preferred, Terraform or CDK experience translates
• Track record building developer platforms: CI/CD, paved-path tooling, internal services that other engineers actually want to use
• Shipped enterprise auth (SSO/SAML/SCIM) and the audit/SOC 2 work that comes with it
• Comfortable in Rust or Python for systems work
• You think of platform as a product, not a service
• You've been the one paged at 3am, and you decided to fix the underlying system, not just the symptom

Technical Stack

• Backend: Rust
• Frontend: TypeScript, React
• Data: PostgreSQL
• Infra: AWS, Pulumi
• AI: Model-agnostic (Claude, GPT)