AI-First Delivery Orchestrator

Y Your Next Career Move Starts Here! The Role: AI-First Delivery Orchestrator Type: Contract What You Ll Do We are looking for an AI-first practitioner who compresses delivery timelines by orchestrating Claude (and other LLMs) across the full development lifecycle, while maintaining the architectural and product judgment that makes the output durable. This person is force-multiplier-shaped: one of them replaces a small team for the right kind of work and produces artifacts that survive enterprise review. How This Person Operates

• Directs and guides AI across planning, solution design, development, documentation, and UAT treating the AI as a capable but literal collaborator that needs clear acceptance criteria to perform well
• Defines what "good" looks like up front (acceptance criteria, conformance checks, architectural fit) and validates AI output against that standard rather than rubber-stamping it
• Owns end-to-end delivery from ideation through production and iteration; acts as product thinker, architect, and delivery lead not just an executor
• Practices context engineering: knows what reference material (code collections, schemas, utility libraries, prior decisions, SOPs) to feed the model, how to scope work so AI has enough context to succeed but not so much it drifts, and how to manage long-running sessions and context windows
• Builds living documentation as a deliverable, not an afterthought Mermaid/BPMN diagrams, structured SOPs, and conformance criteria that make AI output checkable against intent over time
• Breaks work into well-defined, digestible units with clear acceptance criteria
• Validates requirements with business stakeholders to surface and eliminate legacy holdovers rather than encoding them into new systems

Ideal Background

• Hands-on developer history combined with architectural and product ownership; can drop into code to debug, spike, or validate AI output, but doesn't default to hand-coding as the primary delivery mode
• Experience owning product state and requirements, not just implementing tickets
• Comfortable questioning legacy patterns, improving designs rather than replicating them, and pushing back constructively on stakeholder requests that don't hold up
• Has worked inside enterprise environments with real constraints (security, change management, internal source control, data governance) this is not a greenfield startup role

AI Fluency Expectations

• Uses LLMs to perform the majority of development and maintenance work; spends their own time on planning, orchestration, validation, and outcome alignment
• Knows how to be prescriptive with AI preventing it from doing the bare minimum, catching when it confabulates, and structuring prompts and context so output is fit-for-purpose the first time rather than the third
• Exercises judgment about when not to use AI recognizes when a 10-minute hand edit beats a 40-minute prompt-and-validate cycle. Avoids AI ceremony for its own sake.
• Familiar with patterns like CLAUDE.md / project context files, MCP servers and tool integration, structured outputs, and managing parallel or long-running agent sessions
• Claude Certified Architect (CCA) credential is a plus but not required; demonstrated production work matters more than the badge

Tools / Environment

• Required: GitHub (comfortable with enterprise/internal GitHub instances, not just public repos); strong source control discipline
• Strongly preferred: Claude (Desktop, Code, or both); experience configuring MCP servers and skills for enterprise contexts
• Plus: SQL Server / data warehousing background; XML/XQuery exposure or willingness to ramp on MarkLogic; insurance or other regulated-industry experience
• Comfort with mainstream stack tools is assumed (CI/CD, basic cloud literacy, common IDEs)
• Insurance Industry knowledge is a plus

What This Role Is Not

• A traditional data developer role
• A role for someone who primarily writes features or hand-codes solutions
• Transactional use of AI for snippets or speed improvements on otherwise-traditional work
• A pure prompt-engineer role with no engineering depth validation requires the ability to read and reason about code

Why Golden Technology? Founded in 1997, Golden Technology has grown from a two-person vision into a trusted partner for Fortune 500 clients nationwide. Along the way, we ve built a culture centered on family, professional growth, and giving back to our communities through our Golden Community initiatives. Here s What Sets Us Apart •