Principal AI Engineer

About DrFirst

For 25 years, DrFirst has empowered providers and patients to achieve better health through intelligent medication management. We help patients start and stay on therapy with end-to-end solutions that improve prescription access, affordability, and adherence. Our solutions reach 100 million patients a year and are used by more than 420,000 prescribers, 71,000 pharmacies, 270 EHRs, and over 2,000 U.S. hospitals. This is a chance to join a growing Healthcare IT company, work with smart colleagues, and solve complex challenges at a scale few companies match.

Position Overview

DrFirst is building an agentic ecosystem at scale. In the last 90 days we shipped production AI systems across HR, Finance, Legal, and Marketing, automating hundreds of hours of manual work monthly and delivering measurable business impact.

This is a hands-on technical leadership role. You architect and build production AI automation systems from discovery to deployment, working directly with department leaders to map workflows and ship tools that augment their work.

The split is 80% individual-contributor engineering and 20% servant leadership. You build and own technical architecture decisions. As the team scales from 2 to 5-6 engineers over 18 months, you mentor through code review and pairing, teach non-technical leaders to use AI tools, and run discovery sessions that extract workflows from stakeholders. Building stays your primary focus.

Who Will Love This Job

• You are a senior IC who loves shipping code daily. You have been a Staff or Principal Engineer or led small teams, but you miss hands-on building.
• You can architect multi-agent systems and run discovery sessions with department VPs to map their workflows.
• You can debug OAuth2 flows through enterprise proxies and explain security requirements to non-technical stakeholders.
• You are a servant leader who removes blockers, teaches through pairing, and makes stakeholders successful.
• You organize chaos: 15 stakeholders, 6 systems, 3 sprints in flight. You ask "How can I help?" first.

What You Will Work On

Build Production AI Automation (80%)

You own technical implementation of department sprints. Example: an HR sprint delivered resume screening (40-candidate PDF extraction with fuzzy name matching), OKR coaching (Leapsome API integration), performance ratings (calibration detection across 328 employees), and milestone email automation in 4 weeks.

• Architecture: Design and build Python/FastAPI backends connecting to enterprise APIs (HRIS, ERP, CRM, finance, project management).
• Frontend: Build Next.js/React frontends with Okta SSO, role-based access control, and audit trails.
• Multi-Agent Systems: Design agent coordination, skill discovery hubs, agent registries, and workflow orchestration.
• Enterprise Integration: Navigate Zscaler proxies, VPNs, SSL certificates, Kubernetes networking, DynamoDB, and AWS Bedrock.
• Testing: Write Playwright end-to-end tests and maintain a 1000+ test suite with visual regression baselines.
• Deployment: Ship via GitLab CI/CD to AWS EKS (Kubernetes, ArgoCD, Helm, Secrets Manager) and diagnose pod issues.
• Performance: Debug event loop blocking, convert synchronous I/O to async, and optimize CPU-bound operations.
• Security: Partner with the CISO on PII/PHI de-identification, least privilege, and per-user OAuth2.
• Production Support: Debug authentication flows, token expiration, API rate limits, and health check failures.

You will ship 3-5 production tools in parallel each sprint, deploying multiple times per day on stakeholder feedback.

Servant Leadership and Mentorship (20%)

Guide 2-3 engineers through code review and pairing. Teach department leaders to use AI tools. Translate technical decisions for executives. No direct reports and no performance reviews.

• Technical Mentorship: Code reviews, pairing sessions, and unblocking complex integrations.
• Non-Technical Teaching: Run discovery sessions with VPs, teach Finance/HR/Legal leaders to prompt AI tools, and build interfaces they can modify.
• Stakeholder Translation: Explain security trade-offs and translate "OAuth2 delegation" into "everyone sees only their own data."
• Remove Blockers: Clear obstacles, escalate infrastructure issues, and coordinate across departments.
• Knowledge Sharing: Document patterns and write playbooks department heads can follow without engineering help.

Qualifications

Required

• 8+ years building production systems end-to-end (backend, frontend, integrations, infrastructure).
• Staff/Principal Engineer or Tech Lead experience owning technical decisions on production systems.
• Shipped SaaS products with OAuth2, multi-tenancy, audit trails, and enterprise compliance.
• Deep AI/ML experience: production systems using LLM APIs (OpenAI, Anthropic, AWS Bedrock) and agent frameworks.
• Production Python: FastAPI, Pydantic, httpx, pytest (1000+ test suite experience).
• Production TypeScript/React: Next.js App Router, server components, Okta/NextAuth SSO, Playwright.
• AWS and Kubernetes: EKS operations, GitLab CI/CD, ArgoCD, Helm, Secrets Manager.
• Enterprise security patterns: Zscaler proxies, SSL certificates, Kubernetes networking, JWKS verification.
• Healthcare or regulated industry experience (HIPAA, SOC 2, or similar).
• Strong communication: run discovery sessions with non-technical stakeholders, explain trade-offs to executives, write architecture docs leaders can follow.
• Servant leadership mindset and ability to thrive in organized chaos.
• Bias toward shipping: "good enough to get feedback" beats "perfect in 3 months."

Preferred

• Led small teams (2-5 people) through mentorship or tech lead roles.
• US-based (US Central time zone preferred; Poland also considered).
• Built agentic systems: agent coordination, scheduled routines, and feedback loops.
• MCP (Model Context Protocol) experience: custom connectors for enterprise systems.
• Active contributor to AI/ML communities (papers, open source, conference talks).
• Experience teaching non-technical users through workshops and business-user documentation.

AI-Augmented Engineering (Non-Negotiable)

DrFirst runs on AI-assisted development. We use Claude (Anthropic) as our primary AI model via AWS Bedrock, and we are standardizing on AI coding assistants for all development work.

• Active, daily use of AI engineering tools (Claude Code, GitHub Copilot, Cursor, Windsurf, Cline, or similar). Hard requirement.
• Experience with Claude (Anthropic) is highly valued, via Claude.ai, Claude Code CLI, API, or AWS Bedrock. Our platform runs on Claude Sonnet 4.5.
• At least one automation you built that eliminated hours of manual work, ideally using LLM APIs.
• Comfort teaching others to use AI tools through pairing and code review.

Technical Stack

• Backend: Python (FastAPI, Pydantic, httpx, pytest), OAuth2, AWS Secrets Manager, PostgreSQL, AWS Bedrock.
• Frontend: Next.js 14 (React 18, TypeScript, Tailwind CSS), NextAuth.js/Okta SSO, Playwright.
• Infrastructure: AWS EKS (Kubernetes), GitLab CI/CD, ArgoCD, Helm, Zscaler proxies.
• AI and Agentic: Anthropic Claude (AWS Bedrock), AI coding assistants, MCP connectors.

Benefits

• Base salary of $175,000 - $200,000 (exact compensation may vary based on skills and experience), plus a performance-based bonus program.
• Medical, dental, and vision insurance.
• 401K eligible after 3 months, with 50% company match up to the first 5% of salary (3-year vesting).
• HSA with company contribution up to $500 individual / $1000 family per year for HDHP enrollees.
• 100% company-paid short and long-term disability, AD&D, and group life insurance.
• 18 days PTO for the first 3 years of service (increasing thereafter) plus 7 paid holidays.
• Employee Assistance Program and voluntary benefits (FSA, hospital indemnity, accident, critical illness).
• Continuing Education funds up to $1500 annually after 1 year of service.

DrFirst is a Remote-First company. Learn more: https://drfirst.com/company/about-us/careers/

Our recruitment process is secure. You will only be contacted through an official @drfirst.com email address. We will never ask for payment or sensitive personal information at any stage. If a message seems suspicious, verify it at [email protected].