About the role
Own the reference architecture for Halcyra's generative AI deployments across large US health systems. You will lead discovery with CIO and CMIO stakeholders, design secure multi-tenant RAG and agent architectures, set cost and evaluation gates, and stay accountable for the design through build and go-live.
Benefits & perks
Health Insurance
Medical, dental and vision; 100% employee premium covered.
Equity
Meaningful stock options, 4-year vest with 1-year cliff.
Flexible Work
Hybrid: 3 days a week in the Boston office or on-site with customers.
Paid Time Off
20 days PTO plus 11 company holidays.
Learning & Budget
$3,000 annual learning and conference budget.
Retirement Plan
401(k) with 4% company match.
- 01Lead architecture discovery with health-system CIO, CMIO and security stakeholders.
- 02Own the reference architecture for RAG and agentic deployments, including tenancy and PHI isolation.
- 03Model cost, latency and reliability envelopes and set evaluation gates for launch.
- 04Author solution designs and support pre-sales technical evaluations.
- 05Mentor forward-deployed engineers on architecture decisions.
What we evaluate against.
Applied Technical Depth
Must have
- Has designed a production GenAI architecture covering retrieval, tenancy and security.
- Can compare fine-tuning, retrieval and prompting options with evidence.
Nice to have
- Holds a current cloud architecture certification.
Problem Decomposition
Must have
- Has turned a broad AI mandate into a phased roadmap with success metrics.
Nice to have
- Defines kill criteria as well as success criteria.
Customer & Domain Discovery
Must have
- Has led discovery workshops with C-level stakeholders.
Nice to have
- Regulated-industry domain knowledge (healthcare, life sciences).
Product & Solution Judgment
Must have
- Has rejected or narrowed a proposed design or use case on evaluation or cost evidence.
Nice to have
- Has written a governance or model-risk framework adopted by others.
Execution Under Ambiguity
Must have
- Has stayed accountable for an architecture through build and go-live.
Nice to have
- Has supported pre-sales under deadline.
Team Structure
Customer-embedded pods: 2 FDEs, a solutions architect and a clinical informaticist per health-system account.
Operating Rhythm
Monday account reviews, async daily updates, fortnightly deployment retros.
Decision Making
Pods decide scope with the customer; architecture changes go through a weekly design review.
Ambiguity & Pace
Each new health system brings different EHR integrations and security reviews.
Ownership
Pods own a customer deployment end to end, from discovery to go-live.
Customer Obsession
Engineers spend real time inside hospital workflows, not just in meetings.
Transparency
Evaluation results, including failures, are shared with customers.
Learning Culture
Monthly clinical-AI reading group and a generous learning budget.
First 30 days
Review the current reference architecture and deliver a gap analysis against the next three customer deployments.
90 days
Lead architecture for a new enterprise customer from discovery to signed design.
Outcomes
Own a reference architecture that lets every deployment pass security review first time.