Applied AI Specialist, Commercial Customer Success

About the Role Samsara's Commercial Customer Success team supports hundreds of customers who depend on our platform every day to manage their fleets, equipment, and operations. We're building the AI infrastructure to power agentic solutions at scale, and we need an applied AI specialist to design, build, test, deploy, and continuously maintain AI agents that drive value for customers.

What You'll Do

• Build: Design agent architectures, write prompts, integrate with external systems (Salesforce, Samsara platform, Customer Success tools), and write clean Python code that the team can maintain.
• Test: Define success criteria and evaluation frameworks before you write the first line of code. Build automated test suites that cover real-world scenarios, edge cases, and known failure modes.
• Deploy: Own the deployment, versioning, and rollback plan for every agent. Ensure agents are production-ready before they reach CSMs or customers.
• Monitor: Track agent performance in production — monitor accuracy, latency, failure rates, and customer outcomes. Detect drift before anyone has to tell you something's wrong.
• Maintain & Iterate: Update agents continuously as Samsara's product evolves, Customer Success playbooks change, and real-world data indicates what needs improvement.
• Champion, role model, and embed Samsara’s cultural principles (Focus on Customer Success, Build for the Long Term, Adopt a Growth Mindset, Be Inclusive, Win as a Team) as we scale globally.
• You’ll own end-to-end lifecycle of customer-facing AI agents—from concept to production to ongoing iteration—ensuring value delivery for customers and alignment with product and CS playbooks.

In your first 30 days, embed with the Commercial Customer Success team to understand workflows and friction points, get fluent in Samsara platform/data/models, and define success metrics for the first agent. In your first 90 days, ship the first production agentic solution that meaningfully improves customer outcomes and build the monitoring/iteration loop from day one.