Staff Platform Manager, Conversational Products

Airbnb

7d•$200,000 - $240,000•Remote

About The Position

Airbnb’s AI Assistant team owns the agentic AI system powering support for millions of guests and hosts. The Staff Platform Manager will own the platform that determines how the assistant interprets requests, routes them, retrieves knowledge, uses capabilities and actions to resolve issues, and is evaluated for safety and accuracy. The role focuses on increasing the frequency of accurate, complete self-service resolutions while expanding coverage across problem types and user touch points. This is a hands-on product leadership role that combines setting product direction with technical contribution. The manager will diagnose architectural issues, author artifacts that shape production behavior, and guide engineering and data science decisions with partners. They will define successful user outcomes and measurable launch criteria, and align engineering, policy, knowledge, and other cross-functional partners around the product’s direction. The position is US-remote eligible. Occasional work at an Airbnb office or attendance at offsites may be required as agreed with the manager. The employee must reside in a state where Airbnb, Inc. has a registered entity.

Requirements

At least 9 years building technology products, including at least 5 years in product management or a closely related technical role such as engineering, data science, or applied research, with ownership of product direction.
Direct experience shipping generative AI or machine-learning products to production.
Hands-on technical depth, including the ability to read and debug prompts, understand ML and engineering constraints, and author artifacts that shape production behavior.
Ability to work directly with production data, validate analyses end to end, and recognize when results appear incorrect.
Experience designing, calibrating, or certifying automated model-based evaluations and turning vague quality signals into measurable diagnostic components.
Strong sense of user experience, including the effects of latency, tone, and trust on AI interactions.
Ability to own an ambiguous, cross-cutting mandate spanning multiple teams and influence alignment.
Track record of scaling impact through others by setting standards, assigning owners, and maintaining accountability.
Strong executive communication and ability to adapt messages to the audience.
Ability to envision meaningful system improvements and execute a path to achieve them.

Nice To Haves

Experience with retrieval-augmented generation, agentic or function-calling architectures, and evaluation systems.

Responsibilities

Set product direction and a multi-quarter vision for how the assistant reasons and responds, centered on customers and aligned with senior leaders and cross-functional partners.
Work directly with production data, use modern AI tools and coding agents, validate analyses, and investigate results that appear incorrect.
Build new assistant capabilities end to end, from identifying needs through hands-on implementation and calibration with human input.
Define correct and complete resolution outcomes for user problems and align engineering, policy, and knowledge partners around them.
Establish measurable success criteria, including resolution rates, answer accuracy and safety, and launch thresholds.
Own launch readiness for major model and platform changes, balancing speed of learning with required safety and risk work.
Diagnose failures in complex AI systems and determine where fixes should be made.
Own evaluation strategy, including LLM-based evaluators, offline and live-traffic evaluation, calibration, certification, and tools that enable non-engineers to improve the assistant safely.
Align teams with differing perspectives, establish partner agreements early, and promote a single source of truth across the product.
Present regularly to leadership, communicating decisions, tradeoffs, and asks in a way tailored to the audience.
Scale impact through others by setting clear standards, assigning owners, and holding them accountable.