Forward Deployed Engineer

Caylent’s Anthropic Business Unit is a dedicated AI services practice built in partnership with Anthropic. We are Anthropic’s first pure-play delivery partner. 

You are the technical engine of AI engagements. You own the hands-on delivery: rapid prototyping, running evaluations, configuring Claude Code tooling, and building production- grade AI systems in client environments. You live inside your clients’ codebases, standups, and architecture decisions, and you own the technical outcomes. 

There are two FDE archetypes:

• →Own the technical delivery within AI engagements, working alongside the strategy and architecture lead
• →Build rapid prototypes that validate AI approaches early. Prove feasibility before committing to full implementation
• →Design and execute evaluation frameworks: define metrics, run evals against client use cases, and use results to iterate on solutions and inform strategic recommendations
• →Configure and customize Claude Code tooling tailored to each client’s codebase, coding standards, and development practices
• →Design and build production-grade AI systems: agents, data pipelines, AI-augmented workflows, and platform capabilities in the client’s codebase, not a sandbox
• →Own the technical relationship with client engineering teams: you are the primary technical contact from IC to engineering leadership, building the trust that earns you a seat in hard conversations
• →Run technical discovery sessions, architecture reviews, and hands-on workshops with engineering stakeholders; translate strategic direction into scoped, deliverable work
• →Surface technical risks and opportunities. Your field observations shape the engagement strategy

• →Transfer lasting capability: build as if you’re handing off tomorrow, with documentation, runbooks, and working software that stands on its own after you leave
• →Feed the practice: contribute reusable patterns, evaluation templates, and field-tested IP back to the team at engagement close. This is how the business gets smarter from your
  engagements

• Support pre-sales scoping and technical discussions alongside the account team
• Mentor junior engineers and contribute to the delivery methodology
• Collaborate with the engagement team to refine delivery patterns and improve handoffs between strategy and implementation

• 5+ years of software engineering, with meaningful time in a client-facing or embedded delivery context
• Hands-on experience with Claude Code or equivalent AI-assisted development tools, including configuration, prompt-based instructions, and workflow automation.
• Hands-on depth in one or more of: LLM application development, cloud-native architecture (AWS preferred), Kubernetes, or backend systems at scale
• Experience with rapid prototyping. You can take an ambiguous problem and have a working proof-of-concept in days, not weeks
• Familiarity with evaluation frameworks for AI systems: defining metrics, running evals, and using results to iterate on solutions
• Proven ability to run independently inside a client environment, managing your own priorities, stakeholders, and delivery quality while aligning with the broader engagement strategy
• Strong communicator: technically precise with engineers, clear and confident with technical leadership; conceptual and easy to understand for the business; adjusts depth to the audience
• High tolerance for ambiguity; able to start delivering value before every requirement is defined

• Experience building production AI systems: RAG pipelines, agents, fine-tuned models, or LLM-native applications
• Direct experience with Claude Code’s extensibility model: custom skills, hooks, MCP servers, and prompt architecture. Claude Certified Architect Certification preferred.
• Background in a consulting firm, systems integrator, or professional services environment
• Familiarity with SOW-based delivery and scope management in a professional services context
• AWS certifications: Solutions Architect Professional, DevOps Engineer Professional, or equivalent
• Experience designing and running technical workshops or enablement sessions for engineering teams

The FDEs who do well here share a specific profile:

• Earns trust fast. Client engineers invite you into hard conversations because you’ve shown you’ll be straight about trade-offs. Technical credibility and interpersonal judgment both matter.
• Proves it before you build it. You prototype quickly, run evals to validate assumptions, and use data to shape what gets built next. You don’t over-invest before the approach is proven.
• Writes code that lasts. You build as if you’re handing it off tomorrow: clean, documented, transferable. In services, you often are.
• Thinks in systems. You see how today’s technical decision creates tomorrow’s operational constraint. You zoom out before you zoom in.
• High learning rate. T-shaped, with breadth across domains and depth in at least one. Comfortable being new to something on Monday and useful by Wednesday.
• Partners well. Your field observations and technical judgment inform the engagement strategy, and you stay connected to the bigger picture rather than optimizing in isolation.
• Drives to closure. Once a problem is solved, you move to the next priority without dragging prior noise forward. Momentum matters here.

This is a poor fit if you prefer heavy process, close management, or environments where scope is fully defined before you arrive. Ambiguity is the default state here, not a temporary condition.

• Ground floor of something real: this BU is early, well-resourced, and anchored to Anthropic. You’re writing the playbook, not inheriting one
• The work is at the frontier: LLM-native delivery, agents, AI product builds. Not catch-up projects for organizations still evaluating whether AI is real
• Autonomy with backing: minimal oversight in the field, but a strong technical bench behind you when the problem warrants it
• AI-native by default: this team uses AI in how it works, not only in what it ships. An agentic SDLC is how we move from prototype to production fast
• Outcomes, not hours: compensation and advancement tied to delivery performance and revenue contribution, not utilization
• Breadth that compounds: you’ll grow technical range across a diverse client portfolio faster than any single-company role