What are the responsibilities and job description for the AI Solution Architect with Healthcare experience - Dallas, TX position at Photon?
Ideal profile in one line: a hands-on architect (12 15 years) who has shipped LLM/agentic systems in production, has healthcare payer or provider domain scars, has integrated against third-party scheduling or EHR APIs, and can operate as the client-facing technical face of a fixed-bid engagement.
The two hardest-to-find intersections and where I'd focus screening are production agentic-AI experience healthcare compliance fluency, and architecture depth fixed-bid estimation discipline. Plenty of candidates have one side of each pair; this engagement genuinely needs both.
1. Agentic / AI systems architecture (the core differentiator)
- Multi-agent system design this person architects against HAIO master-agent/sub-agent hierarchy and must design scheduling capabilities that plug into an agent platform someone else owns
- LLM application architecture: tool/function calling, orchestration patterns, prompt/context management, evaluation and guardrail design
- Browser-agent automation architecture headless browser fleets (browserless.io specifically), designing agents that generalize to unseen websites, idempotency and duplicate-booking prevention, resilience to site changes and bot controls
- Fallback-chain / graceful-degradation design the four-tier scheduling ladder (API browser agent dialer self-service) is fundamentally a resilience architecture problem
2. Integration architecture
- REST API integration at enterprise scale, webhook/event-driven architectures (the NexHealth confirm/modify/cancel event flows), API gateway patterns (they design against X42)
- EDI familiarity Mukund flagged it as a likely intermediate integration path
- Healthcare interoperability is a strong plus: FHIR, HL7, EHR ecosystem knowledge (NexHealth syncs to EHRs understanding what's underneath the API matters when things break)
3. ML / personalization system design
- Not building the models personally, but architecting the NBA orchestration layer: how propensity signals from pharmacy, health, and MinuteClinic domains resolve into one coherent recommendation
- Understanding of propensity/ranking models, feature pipelines, and event-driven personalization well enough to design contracts between the data layer (Adobe Analytics, member events) and the recommendation layer especially tricky here because that data infrastructure is being built in parallel by another team, so the architecture must tolerate evolving signal sources
4. Mobile & web platform architecture
- Micro-app / modular mobile architecture: Health 100 is native iOS/Android foundations with Flutter micro-apps for new modules the architect must design features that span that seam
- React web architecture for the November website launch
- Push notification, deep-linking, and notification-center integration patterns across channels
5. Healthcare domain & compliance fluency
- HIPAA/PHI data-flow design: data minimization in agent payloads, keeping PHI inside -controlled environments, audit-trail architecture
- Clinical-appropriateness guardrail design knowing that a recommendation engine in healthcare needs approved recommendation classes and prohibited ones, with compliance review gates
- Legal-exposure awareness for automation: T&C permissibility gating for the browser agent, TCPA-adjacent calling conduct for the dialer
- Prior payer/provider/pharmacy domain experience is what separates a good candidate from the right one here
6. Enterprise & delivery skills
- Architecting within someone else's platform estate X42, HAIO, the IVR, and DevOps are all owned by teams that won't grant deep access; this requires an architect comfortable designing to published interfaces and negotiating contracts across team boundaries, not a greenfield architect
- Fixed-bid discipline: decomposing scope into estimable, acceptance-testable milestones their M0 architecture directly underwrites the commercial commitments, so estimation accuracy is a financial skill here
- Stack-ranking and roadmap judgment the "universe of integrations" exercise requires weighing integration difficulty against member value
Salary : $160,000 - $200,000