The company's corporate GenAI group, part of the Office of the CTO, helps its diversified businesses build AI skills and change how work gets done. The group tests cutting-edge capabilities, builds useful systems, develops reusable methods, and helps business teams adopt them. Its scope covers workflow change, product development, AI software factory practices, and AI-assisted cybersecurity.
As Senior Forward Deployed Engineer, you will build AI systems and engineering approaches that drive change across the company's businesses. You will blend product judgment, solid engineering, and experiments with new models to take an unclear problem to dependable use. As a senior individual contributor, you will make architecture decisions, mentor engineers, and build capabilities that reach beyond a single project.
You will work on agent systems, multimodal applications, automation, and AI software factory practices using OpenAI, Anthropic's Claude, and other relevant systems. You will assess new capabilities, set up reusable engineering approaches, and work with security teams on AI-assisted defense.
Responsibilities
Find real workflow needs with business teams, question current assumptions, and prototype solutions that show both business value and reuse potential.
Create, run, test, and maintain reusable agent skills: modular instruction sets for core workflows, bundled with reference material and supporting code. Define when each skill applies, what it depends on, and what permissions it needs. Version and evaluate skills, adapt them to other systems where useful, and help colleagues use them reliably.
Build, review, and test software with AI coding tools. Compare models and tools on output quality, cost, latency, and operational needs. Check delegated work and generated code, and take responsibility for how the system behaves.
Design agent workflows using APIs, tool calls, and Model Context Protocol (MCP). Engineer context, retrieval, memory, and state. Use orchestration and multi-agent approaches only when evaluations show they help.
Build representative evaluations and regression tests covering task outcomes, tool actions, and failure cases. Inspect traces, validate changes, and track quality and cost after release.
Enforce least-privilege access, isolated execution, secrets management, and audit logs, and require human approval for consequential actions. Test for prompt injection and data leakage at the model, tool, and workflow boundaries.
Ship maintainable systems with testing, monitoring, recovery, and rollback. Publish reusable components and coach corporate and business engineers to adopt and support them.
Work with security teams on authorized vulnerability discovery, triage, patch validation, and defensive automation. Help harden systems against more capable AI-enabled attackers, including coordinated multi-agent attacks.
You own the technical design and quality of corporate systems and shared components. Product Owners set priorities, Project Managers coordinate delivery, and Product Ops supports measurement. Work with local technology owners on integration and support, and make your methods usable across the portfolio.
Requirements
Typically 5+ years building software or technical products, or equivalent depth, including independent architecture decisions and ownership of production systems.
Strong engineering fundamentals in APIs, data stores, authentication, cloud environments, testing, and production debugging. You can inspect and correct generated code.
Hands-on experience shipping generative AI workflows or agent systems, including creating and using modular agent skills. You work across AI development tools and verify results as part of daily engineering.
Sound judgment on evaluations, reliability, context and data quality, tool permissions, and human oversight. You can pick a simpler design when it performs better.
Track record of shaping technical direction, mentoring engineers, engaging business leaders, and changing established engineering methods when new evidence supports it.
Experience with Python or TypeScript, agent SDKs or managed runtimes, MCP integrations, multimodal systems, and AI-assisted security is useful but not required. Problem solving and a willingness to keep learning as the field changes matter more than any specific framework or model version.
Experience with computer-use agents that operate browser or desktop interfaces, or always-on agents that keep context and run scheduled or event-driven work, is desired but not required. Relevant engineering experience includes reliable tool execution, monitoring, and recovery from failures.
Strong professional judgment and a readiness to rethink established methods. You question inherited workflows, test what AI can now do, change course when evidence supports it, and teach others. This is assessed through learning behavior and demonstrated work, not a finished playbook on day one.
Conditions
Base salary range: $175,000 to $200,000 USD. This is the company's good-faith estimate for the posted work location at the time of posting. The final starting salary depends on lawful, non-discriminatory factors such as skills, experience, training, certifications, education, location, and market conditions.
Milestones are agreed with your manager and partners based on the work and its readiness.
Expected progress: within 30 days, validate a workflow prototype with explicit evaluation criteria and identify a reusable technical approach. By 60 days, move a solution into a controlled pilot or release with monitoring, access controls, and agreed support. By 90 days, show reliable use and measurable improvement, and enable another team to adopt a tested agent skill or shared component with clear ownership.
Benefits
Competitive benefits package including medical, dental, vision, disability and life insurance, 401(k), paid holidays and paid time off, employee assistance programs, and more.