Staff Software Engineer

About the role

You will architect and own the core infrastructure that powers every AI initiative across the company. You will define the abstractions that allow product teams to build AI features safely and quickly. You will implement the foundational systems that turn high-level product intent into reliable, production-grade AI workflows. You will act as the technical steward of the Agentic Platform, ensuring its longevity and scalability. You will partner directly with executive leadership to align platform strategy with company objectives. You will solve ambiguous problems where the right architecture is not yet clear. You will mentor engineers across the organization on effective platform usage. You will ensure that the platform evolves in step with the rapidly changing AI ecosystem.

Key facts

What you'll do

• Design and build the Agentic Platform SDK using TypeScript, delivering core primitives such as Completion, Agent, Tool, Guardrail, PromptPack, Eval, and Context.
• Implement the Agentic Platform Service that provides Agent-as-a-Service for long-running, asynchronous tasks with robust execution guarantees.
• Create and maintain a prompt management system that supports externalized, versioned prompt storage with seamless CI/CD integration.
• Develop production AI-powered voice and chat applications that serve as the primary proving ground for all platform patterns and capabilities.
• Own provider abstraction layers to support OpenAI, Anthropic, and Google, enabling pluggable backend integrations.
• Engineer structured output validation frameworks that enforce schema compliance and data correctness for LLM responses.
• Build streaming infrastructure that delivers low-latency, reliable data transfer for real-time AI interactions.
• Implement token management systems that track, optimize, and control token usage across all platform interactions.
• Construct safety and compliance infrastructure featuring composable guardrail systems and PII detection and redaction.
• Establish audit logging and privacy-first observability that ensure sensitive data is never exposed to third parties.
• Create evaluation infrastructure capable of systematic quality measurement for non-deterministic LLM outputs and model behaviors.
• Define datasets, scorers, and evaluation strategies including exact match, LLM-as-judge, and schema validation techniques.
• Integrate evaluation into CI/CD pipelines to enable automated regression detection and performance tracking.
• Lead churn containment efforts by designing provider adapters and SDK architectures that absorb frequent model and SDK updates.
• Architect prompt lifecycle management systems with version control, Langfuse integration, and GitHub-based review workflows.
• Design deployment pipelines that automate testing, validation, and promotion of prompt changes.
• Design Agent-as-a-Service infrastructure for long-running async tasks leveraging AWS EventBridge, DynamoDB, and PostgreSQL.
• Collaborate with consuming engineering teams to discover needs, onboard them to the platform, and provide deep technical support.
• Influence architecture and technology selections across the organization by defining and evangelizing platform standards.
• Create reference implementations and comprehensive technical documentation to accelerate platform adoption.
• Champion quality engineering practices including rigorous testing, strict type safety, and end-to-end observability.

Requirements

• Bring 8+ years of software engineering experience with significant time spent building platform infrastructure, developer tools, SDKs, or distributed systems.
• Demonstrate production experience with LLM or AI systems, having built and operated services using OpenAI, Anthropic, or similar providers.
• Show deep understanding of the unique challenges presented by LLM systems such as token limits, non-deterministic outputs, provider outages, and model deprecations.
• Exhibit strong TypeScript expertise, as it is the company standard and you will design APIs consumed by thousands of developers.
• Prove ability to design APIs and abstractions that balance power with simplicity and that are loved by consuming engineers.
• Display a track record of owning complex technical decisions in ambiguous environments while delivering high-quality results.
• Highlight experience partnering with internal customers to align platform capabilities with real-world needs.
• Commit to building stable, maintainable systems that can withstand rapid change and frequent updates from AI providers.

Practical notes

LENGTH: 700-900 words. No HTML, no markdown, no em dashes.

Output the page only.