Senior Director of Engineering, AI Agents

Podium•Lehi, UT

About The Position

Podium is seeking a Senior Director of Engineering to lead the AI Agents product. This role involves owning the technical architecture and roadmap for Podium's AI agent platform, which includes orchestration, tool-use frameworks, memory, and reasoning systems. The successful candidate will lead and grow a team of engineers, build and mature evaluation infrastructure, drive the development of new agent capabilities, and partner with Product, Infrastructure, and customer-facing teams. This is a hands-on leadership position focused on both building new systems and operating them at scale.

Requirements

10+ years of software engineering experience, including significant time in engineering leadership roles (Director, Head of Engineering, or equivalent scope as a hands-on lead)
Direct experience building and operating production LLM-powered or agentic systems — not just prototypes, but systems serving real users at real volume with real reliability requirements
Deep fluency in applied AI architecture: orchestration frameworks, tool-calling/function-calling patterns, RAG and context management, and the tradeoffs between latency, cost, and quality
A track record of building and scaling evaluation systems for non-deterministic software — you know how to measure agent quality, catch regressions, and make shipping decisions without waiting for perfect certainty
Experience leading both "0 to 1" builds and "1 to scale" hardening — you've stood up something from nothing and you've also kept something running under real production load
Strong instincts for working cross-functionally with Product, Infrastructure, and customer-facing teams, and the communication skills to align technical decisions with business outcomes
Genuine, hands-on fluency with modern AI coding tools (Claude Code, Codex, Cursor, or similar) as part of how you and your team build

Nice To Haves

Experience building agent systems specifically for customer-facing, revenue-critical workflows (sales, support, scheduling, or similar)
Familiarity with voice AI, conversational AI, or real-time multimodal systems
Experience operating in a high-growth SaaS or vertical software company, ideally one serving SMB customers
A point of view on the frontier of agentic AI — what's changing fast, what's overhyped, and where the real leverage is for a company like Podium

Responsibilities

Owning the technical architecture and roadmap for Podium's AI agent platform — the orchestration layer, tool-use frameworks, memory, and reasoning systems that power customer-facing agents across every industry vertical we serve
Leading and growing a team of engineers building and operating production agent systems, setting the technical bar and staying hands-on in design, code, and incident response where it matters most
Building and maturing our evaluation infrastructure — the offline and online eval harnesses, quality metrics, and feedback loops that let us ship agent improvements with confidence instead of guesswork
Driving the "0 to 1" build-out of new agent capabilities and workflows, then the "1 to scale" work of hardening those systems for reliability, latency, and cost as usage grows into the tens of millions of conversations
Partnering closely with Product to translate business outcomes (response time, conversion, show rate, revenue recovered) into agent capabilities and prioritized technical bets
Partnering with Infrastructure/Platform teams to ensure agent systems are observable, cost-efficient, and resilient at scale — this includes model routing, latency budgets, fallback behavior, and guardrails
Working directly with customers and customer-facing teams (Sales, CS, Support) to understand where agents succeed or fail in the field, and feeding that signal back into the roadmap
Establishing best practices for prompt/context engineering, agent testing, safety and guardrails, and responsible deployment of autonomous systems in live customer environments
Representing AI Agents engineering in leadership conversations — setting a clear technical narrative, calling out risk early, and making sound build-vs-buy and model-vs-vendor tradeoffs