Product Designer

Meta; New York (NY); On-site; USD 201,000 - 278,000; Full time

Job summary

Meta’s Business Agent team is seeking a Product Designer in New York, NY (onsite) to shape high-quality agent behavior across chat, voice, and visual surfaces.

You will write and iterate prompts, test surfaces, and collaborate with engineering and data science to define evaluation strategies and quality metrics for multi-modal agent experiences. This role emphasizes measurable impact and cross-functional collaboration.

Qualifications

• Bachelor’s Degree in HCI, design, or related field or equivalent work experience.
• 11+ years designing software products with focus on UX, interaction, and visual design.
• Experience building and maintaining evaluation pipelines for LLM products, including graders, rubrics, and regression suites.

Responsibilities

• Write and iterate prompts that power Meta’s agentic tools across a product surface.
• Test the surface, translate findings into prompt changes, ship updates, and confirm users receive the intended prompting.
• Collaborate with engineering and data science to define quality metrics and evaluation strategies for multi-modal agent outputs.
• Define, evolve, and unify visual and interaction systems that scale across products and platforms.
• Present design strategy and rationale to leadership and cross-functional executives.

Skills

UX design

Product design

Interaction design

Visual design

Prototyping tools

Education

Bachelor’s Degree in HCI or related field

Tools

LLM evaluation tools

Prototyping tools

Job description

Meta’s Business Agent team is building multi-modal AI agent experiences end to end, from early prototyping to prompt design and evaluation. As a Product Designer in New York, NY (onsite), you will shape what high-quality agent behavior looks like and ensure it works reliably across chat, voice, and visual surfaces.

What you’ll do

• Write and iterate the prompts that power Meta’s agentic tools, features, and behaviors across a product surface
• Test the surface, translate findings into prompt changes, ship updates, and confirm users receive the intended prompting
• Package what prompting alone cannot solve for training, including graders for crisp behaviors and human-feedback questions with good/bad pairs for fuzzy outcomes such as writing quality
• Collaborate with engineering and data science to define what good looks like in agent output across chat, voice, and visual experiences, and drive an evaluation strategy to measure it
• Extend evaluation coverage beyond text so voice and visual agent outputs have clear quality judgments tied to the conversational or visual nature of each surface
• Define quality bars that scale across a long tail of businesses, supporting a single agent specification from small business through enterprise without bespoke rework
• Rapidly prototype new ideas and experiences, helping the team prioritize upcoming workstreams based on impact and feasibility
• Define, evolve, and unify visual and interaction systems that scale across products and platforms
• Define evaluation strategy and build evals, golden sets, judges, and related components independently and alongside data science and engineering
• Apply familiarity with how code works to update language and prompts within internal tools, review changes to verify the shipped experience matches intent, and build logic that adapts agent behavior to context
• Work with product and cross-functional partners to set a strategy for multi-modal agent communication
• Present design strategy and rationale to leadership and cross-functional executives, helping align organizational priorities across product, engineering, research, and content strategy

What you bring

• Bachelor’s Degree in HCI, design, or related fields, or equivalent work experience designing and building agentic experiences
• 11+ years designing software products with a focus on user experience, interaction, and visual design aligned to business objectives
• Experience building and maintaining evaluation pipelines for LLM products, including graders, rubrics, comparison sets, regression suites, and the systems that run them across models
• Experience designing and building agentic experiences
• B2C or B2B experience
• Deep expertise in prototyping tools and the ability to evaluate and advance design craft across varying fidelity levels

Nice to have

• Experience implementing responsible, ethical AI practices, such as risk assessment, bias mitigation, and quality and accuracy reviews
• Proven ability to integrate AI tools to optimize or redesign workflows and deliver measurable impact, including efficiency gains and quality improvements
• Experience converting product rubrics into training signals, such as graders, human-feedback questions, or preference pairs
• Experience shaping agent behavior on voice or other multi-modal surfaces
• Ongoing AI skill development, including prompt/context engineering and agent orchestration, with staying current on emerging AI technologies
• AI-native design fluency, including experience with model design, generative AI logic, and defining assistant behavior and tone at the systems layer
• Front-end or notebook-to-app experience, and opinions about what makes an eval result legible at a glance
• Experience with both B2C and B2B is preferred

Compensation: USD 201,000 to 278,000 per year.