Senior AI Software Engineer, Agentic Systems & Data Fabric
Careers Kinaxis, Remote, CA
The role
The Data Fabric team is a cross-functional team at Kinaxis that builds a multi-tenant, multi-cloud data platform on Databricks (GCP & Azure) to connect, govern, secure, and discover data for supply chain and AI products. We deliver secure tenant isolation, governed data sharing, and the harness enablement layer that lets agentic AI safely reason about and act on enterprise data. As a Senior AI Software Engineer on the Data Fabric team, you will design and implement execution patterns, tool-calling infrastructure, and runtime guardrails that enable agentic workflows to operate reliably on enterprise-grade data at scale.
Vacancy Status
This is an existing job vacancy
What you will do
- AI harness & execution layer: Build core services and reusable runtimes that connect LLMs and agents to governed data on Databricks.
- Data Fabric integration: Wire agent workflows into Delta Lake, Unity Catalog, SQL endpoints, and multi-cloud pipelines across GCP and Azure.
- Tenant isolation & data safety: Design multi-tenant boundaries across data, compute, identity, and execution context.
- Platform components: Ship production-ready APIs, SDKs, reference implementations, and runbooks that teams adopt at scale.
What you need
- Workflow orchestration & system design: Track record architecting resilient execution workflows, APIs, and production platforms.
- Agentic engineering: Hands-on production experience with agent frameworks, tool calling, state management, approvals, retries, and evaluation.
- Closed-loop reliability: Experience designing observable, self-correcting workflows with planning, independent verification, and human governance.
- Testing & guardrails: Experience applying automated testing, runtime feedback loops, and guardrails to validate agent-generated outcomes.
- Python development: Deep proficiency writing scalable Python runtimes, asynchronous services, and developer SDKs.
- Strong initiative: Proactive builder who independently turns ambiguous problem spaces and client needs into impactful platform products.
Nice to have
- Databricks ecosystem: Hands-on experience with Unity Catalog, Delta Lake, and SQL endpoints on GCP or Azure.
- Infrastructure as Code:
