Dynamic Bay Area startup is seeking an AI Interface Engineer who will own the design and development of agent-driven interfaces across both our customer-facing product and internal tooling. This person will be a founding voice in shaping how to approach AI-native frontend development – building a workflow where early prototypes run against live platform data via MCP and mature quickly into shippable experiences. The core of this work spans conversational AI interfaces, agentic workflows, and the MCP servers that connect agents to the real-time data platform relied on by customers.
Most Important Responsibilities
Work closely with product, design, and engineering to create a repeatable path from prototype to production, where agent-mediated experiences are tested against live data over MCP before they ever reach a spring review.
Build conversational and agentic experiences directly into our product surface, giving customers natural language access to the workflows and data that underpin our real-time platform.
Develop and maintain product-facing MCP servers and agent tooling and internal automation – safely and at scale.
Apply these same agent-driven patterns to internal tools first – configuration interfaces, admin workflows – proving out the approach before rolling it outward to customer-facing surfaces.
Own prompt engineering end-to-end: authoring, versioning, and systematically evaluating prompts, tool specs, and agent behaviors using reproductible test harnesses tied to measurable outcomes.
Build out an evaluation framework for agentic UX that tracks task completion, latency, guardrail adherence, and continuous improvement loops grounded in real usage data.
Champion agent-mediated interface patterns organization-wide, working across frontend, backend, product and platform to make this a core architectural approach rather than an afterthought.
See initiatives through from initial discovery and prototyping all the way to production launch, instrumentation, and iterative improvement.
What you can bring to the table to impact this role, team, and organization:
Demonstrated experience shipping production applications built around LLMs and agent frameworks, including tool use, multi-turn context management, and safe handling of non-deterministic model outputs.
Strong product and UX sensibility: ability to translate ambiguous workflows into clear interaction designs and to advocate effectively for user needs in cross-functional discussions.
Fluency in prompt engineering, including multi-turn conversation design, tool definition, context-window management, and systematic evaluation behavior.
Full-stack engineering ability, with a strong frontend depth in Typescript and React and the ability to move into Python backends, built REST/WebSocket services, and wire interfaces to real data sources.
Hands-on experience with agent tool protocols (MCP or equivalent) and with designing tool surfaces that are both expressive for agents and safe at production scale.
Experience with agent-mediated code generation tools (Claude Code, Codex) in a production setting, including writing prompts, skills, or agents that non-engineers use.
Experience with cloud platforms (AWS preferred) and using modern CICD workflows.
Comfort operating in ambiguity within an emerging engineering practice, partnering with other founding contributors to set direction and establish patterns across teams.
Preferred Qualifications
Bachelor’s degree in computer science, Human-Computer Interaction, Design, or related field.
Experience integrating with collaboration and knowledge based systems such as GitHub, Confluence, Jira, or Notion, including building programmatic access layers over them.
Experience working directly with designers in Figma or comparable tooling, including translating design intent into working code and feeding design integration from production data.
Background in developer tooling, internal admin platforms, or configuration-heavy products.
Bonus Experience
Experience designing information architecture for knowledge-dense domains, unifying structured and unstructured content (code, documents, issue trackers, wikis) into coherent representations that both humans and agents can navigate.
Experience with retrieval, semantic search, or vector-store backed knowledge systems at sclae.
Experience building and operating microservices into production.
Exposure to IoT device management, real-time data systems, or other data-dense domains where agent-mediated interfaces offer clear leverage.
Featured Benefits: Medical, Vision, Dental, Stock Options