About the Role
Carrot Fertility is seeking an Applied AI Engineer to join our Enterprise Technology team. This role involves designing, building, and deploying production-grade AI and integration solutions to provide internal teams with reliable, structured access to Carrot's core product and operational data. It is a hands-on engineering position where you will be responsible for end-to-end delivery, from scoping and architecture to deployment, iteration, and demonstrating measurable business impact.
Your initial project will focus on establishing the data access layer for Carrot's enterprise AI agent ecosystem. This involves designing and deploying an MCP (Model Context Protocol) architecture to expose structured, governed access to critical operational data such as member eligibility, benefit balances, expense records, provider information, and employer-specific rules. This foundational infrastructure work will be a regular part of your responsibilities as Carrot's AI capabilities expand. You will work closely with internal teams across Operations, Business Systems, and Product, translating their data access requirements and workflow challenges into AI-powered solutions that provide lasting operational efficiency. The role demands the urgency of a startup engineer combined with the judgment of a senior architect, while adhering to a core design principle of least complexity: building the right solution with the appropriate tools in a maintainable and extensible manner.
What You'll Do
- Embed directly with internal business teams to identify data access gaps and workflow pain points, rapidly prototype solutions, and manage the complete delivery lifecycle from scoping to production deployment.
- Architect and construct agentic AI systems capable of handling complex, multi-step business processes, ensuring reliable, deterministic, and auditable outcomes, particularly in high-stakes or regulated environments.
- Design systems with integrated compliance and data governance, including HIPAA-compliant data handling, role-based access control, prompt hygiene, evaluation frameworks, and comprehensive observability.
- Write high-quality, production-grade code and develop platform-based integrations, selecting the most suitable tool for each task. This includes utilizing low-code solutions where they reduce delivery time and maintenance, and custom code when precise control or unique capabilities are required.
- Utilize Claude Code as your primary AI-assisted development environment, deeply leveraging agentic coding to accelerate delivery and maintain high output quality.
- Translate ambiguous business requirements into clear technical designs and effectively communicate them to both technical and non-technical stakeholders.
- Define and instrument success metrics for each solution, such as time saved, error rates, SLA improvements, cost reduction, and user satisfaction.
- Document patterns, failure modes, and reusable frameworks to contribute to a shared internal playbook, enhancing team velocity over time.
- Lead enablement efforts, conducting knowledge-transfer sessions and creating documentation to ensure business teams can understand, trust, and extend the solutions you build.
- Build and maintain MCP (Model Context Protocol) servers that expose Carrot's core product and operational data to AI tools and agents across multiple data domains.
- Develop a deep familiarity with Carrot's application database schema to independently navigate the data model, design reliable queries, and create clean access patterns without continuous support from Product Engineering.
- Develop Claude skills and plugins that business teams and AI agents can utilize, extend, and maintain.
- Apply a design principle of least complexity to every build, preferring simple, maintainable, and documented solutions over sophisticated ones, and avoiding the creation of single-person knowledge dependencies.
Example Projects You Might Tackle
- Application data access layer and AI agent enablement (your first project): Design and deploy a suite of MCP servers that provide structured, governed access to Carrot's core product and operational data. This includes data domains such as expense records, uploaded files, member eligibility and benefit dates, benefit balances, merchant and currency data, journey and phase, medical and clinical questions, care recipients, employer-specific rules, and pharmacy network access. Integrate this MCP layer into Carrot's AI agent infrastructure and fine-tune the agent to significantly improve the quality and reliability of AI-assisted internal operations.
- AI-assisted intake and routing: Develop an intelligent triage system that classifies inbound requests, summarizes context, and routes or drafts responses for internal operations and customer care teams.
- Revenue and finance automation: Collaborate with RevOps and Finance to create automations that reduce manual data entry, enhance forecast accuracy, and accelerate insight generation.
- RAG-powered knowledge assistant: Deploy a retrieval-augmented system that surfaces authoritative documentation, SOPs, and institutional knowledge within the tools teams regularly use.
- Automated KPI dashboards: Instrument and visualize business impact, including time saved, reduced error rates, and improved SLAs, across all active AI solutions.
About You
- 6–10 years of software engineering experience with strong fundamentals in data structures, system design, SQL, and application integration. Experience in a Business Systems, Enterprise engineering, or internal-facing engineering context is a significant plus.
- Proven track record of shipping production software, with a focus on owning reliability, observability, and the operational health of your creations.
- Highly proficient with Claude Code, utilizing AI-assisted development as a fundamental part of your workflow.
- Hands-on experience building and deploying agentic AI systems, including LLM orchestration, tool/function calling, RAG pipelines, and MCP server development. You possess an understanding of agent reliability patterns and evaluation frameworks.
- Programming fluency in Python and TypeScript/JavaScript; comfort with Ruby is a plus.