Principal Scientist - Agentic Product

The Opportunity

Adobe is reimagining the future of creativity - where generative AI, intelligent agents, and human imagination work hand in hand. As a Principal Scientist (Engineering), Agentic Product, you will be a key technical leader for Adobe's up-and-coming AI-native experience that brings generative models, creative intelligence, and decades of differentiated Adobe capabilities together for hundreds of millions of creators. You will guide the architecture and hands-on buildout of the agent across the full stack and across Adobe's target platforms, from desktop and mobile to web and embedded surfaces. You will help reshape the way the team works in the age of agentic engineering. And you will be getting in on the ground floor to help form the team and its practices.

This is a rare opportunity to define the engineering foundations of Adobe's soon-to-be most visible AI product - from model integration and agentic reasoning patterns, to production infrastructure and cross-platform delivery. The right person brings deep, hands-on expertise with LLMs and agentic systems, a genuine enthusiasm for transforming how engineering itself is done with AI tools like Claude Code and Codex, and the technical depth to make hard calls that hold up over time.

What you'll do

• Guide the end-to-end technical architecture of Adobe's Agentic Product across the full stack - from LLM integration and agentic reasoning layers, to API surfaces, client runtimes, and platform-specific delivery on desktop, mobile, and web.

• Lead the production buildout of the agent: translating prototype-quality capabilities into robust, scalable, maintainable systems that meet Adobe's quality and reliability bar for a large, diverse audience.

• Drive hands-on implementation and technical decision-making for the highest-leverage, highest-complexity components - including model orchestration, tool use, multi-step task execution, context and memory management, and agent evaluation infrastructure.

• Champion the adoption of AI-assisted engineering practices - using tools like Claude Code, Codex, and emerging agentic development platforms to accelerate implementation, raise code quality, and fundamentally change how the team ships. Guide engineers toward new workflows where AI is a first-class collaborator in the development process.

• Guide best practices in developer experience: tooling, local development loops, CI/CD ergonomics, and the scaffolding that lets a team move fast and confidently in a fast-moving AI product environment.

• Stay current with the model capabilities ecosystem, evaluating new LLM approaches and frameworks, and advising the team on when and how to adopt them.

• Work with the evaluation and quality framework for agent behavior - the evals, benchmarks, and systematic feedback loops that allow the team to measure progress and catch regressions as capabilities evolve.

• Partner with Engineering and Product leadership to translate product ambitions into technically grounded delivery plans - surfacing constraints early, proposing architectural options, and ensuring scope is set based on accurate technical understanding.

• Collaborate across Adobe's platform, infrastructure, data, and security teams to ensure the Agentic Product is built on a foundation that supports guardrails, data protection, observability, and the operational requirements of a product at Adobe's scale.

• Mentor and grow senior and staff engineers on the team, establishing technical standards and elevating the team's collective capability in agentic engineering.

What you'll need to succeed

• Deep, hands-on expertise with large language models - including practical fluency with LLM APIs, prompt engineering at scale, fine-tuning and RLHF concepts, context window management, retrieval-augmented generation, and the operational realities of running LLM-backed systems in production.

• Proven experience building and shipping agentic systems: tool use, multi-step reasoning, agent orchestration frameworks, memory and context persistence, and the failure modes that emerge at each layer.

• Hands-on experience with AI-assisted engineering tools - Claude Code, Codex, Cursor, or similar - and a track record of integrating these into real development workflows. Enthusiasm for pushing the boundary of what AI-augmented engineering can look like at the frontier.

• Full-stack engineering depth sufficient to guide architecture and lead implementation across the agent's entire surface area - including backend services, client SDKs, and platform-specific integration patterns for Adobe's target platforms.

• Passion for shaping developer experience: opinionated views on tooling, local dev ergonomics, onboarding, and the scaffolding that helps a team stay productive and confident as a codebase grows.

• Strong communication skills to serve as a credible technical voice alongside PM leads and Adobe's senior technical leadership - able to make hard architectural calls, defend them clearly, and adapt when new information warrants it.

• Comfort operating in a fast-moving, ambiguous environment where the underlying models, frameworks, and user expectations are all rapidly evolving - strong enough foundations to move fast without building on sand.

• Experience collaborating cross-functionally with product, design, platform, and security teams in the context of shipping user-facing AI products.

• Familiarity with the quality, trust, and reliability requirements of shipping AI features to a large consumer and enterprise audience, including responsible AI considerations specific to generative and agentic systems.

• A track record of technical mentorship and raising the engineering bar on teams working at the frontier of applied AI.