Forward Deployed Engineer - Applied AI

About the role

Forward Deployed Engineer, Applied AI

Medicat builds MedicatOne, the EHR used by college health and counseling centers across the US. Banyan Software acquires and grows vertical software companies like ours, and we're investing in AI that reaches clinicians and students in production.

This is an engineering role first. You will write production code most days: LLM integrations, automations, data workflows and the APIs that connect them to MedicatOne. You'll also sit close to our clinic customers and internal teams, turning what you see in their workflows into working software, often in days, not quarters.

If you're happiest owning a roadmap and writing specs for others to build, this isn't the role. If you want to be the person who ships the feature, debugs it with the customer on the call, and hardens it for HIPAA, keep reading.

What you'll do

Expect roughly 70% hands-on engineering and 30% embedded with customers and internal teams.

• Build and ship AI features end to end. Design, code, test and deploy LLM-powered features in MedicatOne: prompt design, tool calling, retrieval where it earns its place, and evaluation harnesses.
• Pick the simplest reliable solution. Sometimes that's an LLM, often it's plain code.
• Integrate with real clinical data. Write the services, SQL and data pipelines that move PHI safely between MedicatOne, customer systems and model providers.
• Deploy alongside customers. Join clinic implementation and discovery calls as the technical lead, scope what can be built, prototype against their data, and stay through go-live.
• Embed with internal teams. Find and automate repetitive work, and help analysts and developers get the most out of AI tools.
• Own production quality. Instrument, monitor and debug what you ship: latency, cost, accuracy and failure modes. Write evals so regressions are caught before customers see them.
• Make AI safe for healthcare. Implement HIPAA-compliant patterns for PHI handling, logging, access control, de-identification and vendor data boundaries.
• Harden internal tools. Review and productionize AI tools built by non-engineering teams before they touch customer data.
• Feed the roadmap with evidence. Bring back what you learned in the field as working prototypes and technical recommendations, not slide decks.

What you bring

• 3+ years as a software engineer shipping production code in B2B SaaS, with an individual-contributor track record. You still write code every week.
• Strong in at least one backend language (Python, C#/.NET or TypeScript/Node) and comfortable reading the others.
• Hands-on LLM API experience: you've shipped production features on LLM APIs (OpenAI, Anthropic, Azure OpenAI), including prompt design and function/tool calling, and you know when not to use one.
• End-to-end system design: you've designed a production system from the ground up, including data flow, security and multi-tenancy.
• Solid data skills: production SQL, schema design and data pipelines against relational databases.
• API and integration experience: REST API design, authentication, webhooks and third-party integrations.
• Production engineering habits: Git, code review, automated testing, CI/CD, and deploying to a major cloud (Azure or AWS).
• Regulated or sensitive data experience: you've built software that handles PHI, PII, financial or similarly protected data, with the access controls and audit trails it requires.
• Comfort with customers and colleagues: you can explain a technical trade-off to a clinic director or an analyst, then go build it.
• Authorized to work in and resides in the US; we can't sponsor visas.

Nice to have

• RAG, embeddings and vector databases (pgvector, Pinecone or similar), plus the judgment to know when retrieval is worth the complexity
• HL7 v2, FHIR or EHR integration experience (we can teach this)
• HIPAA experience specifically
• Prior Forward Deployed, Solutions Engineering or implementation-engineering experience at a software company
• LLM evaluation and observability tooling (e.g., LangSmith, Langfuse, custom eval suites)
• Agent frameworks or the Model Context Protocol (MCP)
• College health, student health or behavioral health software
• Open-source contributions or a public portfolio of shipped AI projects