Principal AI Engineer

Description

The company builds software for in-home health and hospice agencies throughout the US. Its platform covers scheduling, clinical documentation, coding, compliance and revenue cycle, and handles millions of patient visits a year.

The team is building an AI software factory, and this role creates the first working version of its agent platform. Agents will work against the production codebase while people write specifications and check results. The role also acts as the senior technical authority on AI inside the products.

The core stack is.NET and Azure, with Azure DevOps, Kubernetes, Azure AI Foundry and MCP tooling, plus other languages. The platform has to work across all of them. The effort is greenfield, so you set the standards.

This is a hands-on individual contributor role with no direct reports at the start. Expect to code most days, especially in the early months. You shape standards and the roadmap through working code and reference implementations, not decks or committees. It suits someone whose recent work was building systems rather than mainly leading teams or advising.

Responsibilities

• Build and run the agent platform that works autonomously on the production codebase, moving work from specification to pull request with human review.
• Write agentic patterns, orchestration and tool integrations for real engineering workflows.
• Make the platform effective across the whole technology stack, not just one language or framework.
• Set the agent templates and shared framework every team will use.
• Manage the AI lifecycle from concept through deployment, operation and iteration on Azure and Azure AI Foundry.
• Define how work reaches agents and review their output from specification to pull request.
• Work with application teams to add AI to existing products and workflows.
• Build APIs, SDKs, MCP servers and extension points for engineering teams to use.
• Set up packaging, versioning, discovery and distribution for MCP servers, tools and agent skills, such as an internal registry or catalog with access controls.
• Design tool access so an agent working for one agency can only reach that agency's data.
• Design context, skills and retrieval patterns for large codebases, including an application-level knowledge map showing which module a ticket belongs to and what it affects.
• Decide when retrieval fits better than curated context.
• Set versioning, review and governance for shared AI assets.
• Build evaluation frameworks that measure output quality and catch regressions.
• Define guardrails, failure handling and human-in-the-loop controls for regulated settings, keeping PHI out of agent context, prompts and logs.
• Tune AI systems for reliability, accuracy and cost using token budgets, rate limits, retry caps and step limits.
• Set observability practices for supervised and autonomous AI use.
• Mentor engineers company-wide and raise AI skill by building working examples in their own codebases.
• Define engineering standards and shape the AI roadmap with the CTO.
• Test emerging LLM, agentic AI, RAG and MCP capabilities and move proven ones into production.
• Advise on build, buy and partner choices across the AI toolchain.
• Track commercial and open-weight models, and match model tiers to tasks by cost, latency, quality and data-handling terms.
• In the first 90 days, build an agent template and reusable framework that teams can adopt, with at least one team running it on live tickets.
• Add telemetry for token spend, agent launch frequency and traceable tool calls.
• Set a baseline to measure the effect on development: time saved, delivery speed and cost.
• Create one shared standard across the clinical, revenue cycle and integration teams. Each team owns its knowledge, but no team builds its own agents, models or tiers.

Requirements

• You hold a bachelor's or advanced degree in computer science, engineering or a related field, or have equivalent practical experience.
• You have 10+ years in software engineering and write code today, not only review it.
• You have 3+ years building and running production AI systems where agents call tools or change code. RAG or machine learning products alone do not count.
• You are proficient in Python and modern AI tooling such as LangChain, LangGraph, MCP, and the OpenAI and Anthropic SDKs.
• You have built MCP servers or similar tool interfaces, including authentication and per-tenant access control.
• You can make agents work across several languages and technology stacks.
• You have deployed and operated AI workloads in cloud environments, with Azure preferred.
• You understand LLMs, generative AI, agentic patterns and context engineering.
• You have proven RAG experience and know when retrieval beats curated context.
• You have worked with AI evaluation and observability, including quality measurement and cost and token management.
• You know how to handle regulated data such as PHI in AI systems.
• You have a record of mentoring engineers and raising technical skill across teams.
• You can share a repository, code sample or demo of agent or MCP work you built yourself.
• Preferred: experience integrating AI into C and.NET applications.
• Preferred: experience using AI coding tools such as Cursor or Claude Code on large codebases.
• Preferred: familiarity with vector databases and retrieval tools such as pgvector, Pinecone or Azure AI Search.
• Preferred: exposure to LLMOps and evaluation tools such as Azure AI Foundry, MLflow, LangSmith and CI/CD.
• Preferred: familiarity with agent-to-agent standards such as A2A, and knowing when a multi-agent setup outperforms one agent that uses tools.
• Preferred: experience with large legacy codebases and incremental modernization.
• Preferred: interest in healthcare, EHR or revenue cycle applications.

Conditions

The position does not offer immigration sponsorship. Candidates must have the right to work in the U.S. without sponsorship.

Benefits

• Competitive pay and robust benefits.
• Professional development opportunities.
• Work-life balance, with career opportunities designed to fit changing life demands.
• The company supports charitable initiatives each year.