Two-Person IT Team Runs HIPAA Compliance for 400 Employees Across 25 States

#Sprinter Health's $55M Series B Fuels AI Automation Hiring to Scale In-Home Care Across 25 States

The Build-Out Trigger

Sprinter Health closed a $55 million Series B on May 15, 2025, TechCrunch reported. General Catalyst led; Andreessen Horowitz Bio + Health, the Regents of the University of California, Google Ventures, and Accel followed. The round lifts total funding to $125 million for the Menlo Park company Max Cohen and Cameron Behar founded in 2021 after stints at Oculus and Google, TechCrunch's data shows.

The capital targets headcount and geography. Cohen told MedCity News the company will grow clinical, engineering, AI, and operations teams while expanding from 18 states to 22 by summer's end. Job board data shows the flywheel spinning: 14 roles posted in the past week, including AI Automation Team positions at staff ($220K–$255K) and senior ($185K–$230K) levels in San Francisco.

Holly Maloney, who led the seed round and sits on the board for General Catalyst, framed the bet on infrastructure: "They've built the technological infrastructure to make care both scalable and impactful for the people who need it most." That infrastructure, a proprietary routing and engagement platform letting each field clinician serve up to 12 patients daily, is what new engineering hires must scale across state regulations, real-time coordination demands, and HIPAA-compliant data pipelines.

Inside the AI Automation Team's Mandate

The AI Automation Team builds the operational nervous system powering 130,000 in-home visits across 22 states — soon 25. Job postings for senior and staff engineers describe the same charter: own the backend services and APIs that drive patient booking, clinician routing, logistics, and medical device integrations. These systems decide which clinician goes where, when, and with what equipment, then track the visit in real time.

The scope splits into two connected layers. First, pure logistics orchestration: scheduling, dispatch, routing, and safety checks at scale. Descriptions cite "complex operational challenges like scheduling, dispatch, safety checks, and error prevention at scale" and "supply/demand dynamics, including forecasting, routing, scheduling, or inventory management." This is marketplace mechanics applied to healthcare — the class of problem DoorDash or Uber solves, but with HIPAA constraints, clinical protocols, and a 92 NPS to protect.

Second, clinical decision support baked into the workflow. Engineers integrate with external health systems, telemedicine platforms, and medical devices "in a reliable, secure way." They build data flows letting a virtual clinician review vitals collected in a patient's living room, flag preventive-care gaps, and trigger follow-ups. The Applied AI role, listed separately as "Automated Intelligence," extends this with LLM-powered clinical visit summaries, patient risk models, and voice coordination assistants that automate scheduling and follow-up tasks. But the AI Automation Team owns the production rails those models run on: the APIs, the event streams, the device integrations, the audit trails.

Both roles require shipping 0→1 in a regulated environment. The stack (serverless AWS (AppSync, DynamoDB, Lambda, Amplify, CloudFormation, Node), React Native for Web, GraphQL, TypeScript) is chosen for speed and compliance, not novelty. Engineers sit alongside product, data, ops, and clinical teams, turning "real-world problems into shipped software." The mandate: own projects end to end, make architectural decisions, evolve engineering practices as the network grows from 22 to 25 states.

Hiring Profile: Senior and Staff Engineers for 0→1 Clinical AI

Sprinter Health is building its Automated Intelligence team from zero. These are production AI engineering positions shipping clinical summaries, risk models, voice coordination assistants, and workflow agents that automate scheduling and follow-ups — not "AI adjacent" roles. The postings set the bar: you have shipped AI products, not prototypes. You have operated in 0→1 environments with high ownership. You understand reliability, iteration speed, and production safety.

Roles and Experience Bars

Source: First-party board data for AI Automation Senior and Staff bands; Dice posting for the Applied AI Lead variant. The Ladders listing shows $130K–$180K but appears to be an aggregate estimate — the board's live bands are the stronger signal.

The AI Automation postings converge on 8+ years building and scaling production backend or full-stack systems, experience mentoring engineers, and designing APIs, data models, and services that power user-facing products. The Applied AI posting adds hands-on experience with LLMs, RAG pipelines, embeddings, and applied NLP, plus a track record of navigating ambiguity with tight feedback loops. Bonus signals include FHIR/HL7 familiarity, voice agent or telephony API work, and experimentation tooling such as Statsig.

Location Model

Hybrid, Bay Area only. Offices in San Francisco and Menlo Park. Postings describe flexibility for work-life balance but require in-person presence — no remote option. The interview process includes an onsite with a three-hour block: systems design (AI-focused for Applied AI), behavioral, and lunch with the team.

Technical Stack You'll Own

• AI & Modeling: LLMs, RAG pipelines, embeddings; LangChain / LlamaIndex; Python
• Backend & Services: Node.js / TypeScript; GraphQL / REST APIs; AWS Lambda, AppSync, DynamoDB
• Data & Infrastructure: BigQuery, Elasticsearch / OpenSearch; Airflow orchestration; vector stores and retrieval systems
• Bonus Domains: FHIR, HL7, clinical notes; voice agents / telephony APIs; Statsig or internal experimentation tooling

What Differentiates Candidates

Five "edge" criteria filter for the Series B build-out:

1. Shipped AI products, not research demos or prototypes
2. Built systems interfacing with real users or complex workflows
3. Operated in early-stage or the previously noted high-ownership 0→1 environments
4. Understand the previously noted reliability, iteration speed, and production safety
5. Motivated by solving real problems for real people, not AI hype

Compensation Structure

Base salary bands above. Equity is "meaningful pre-IPO equity" across all postings. Benefits: 100% medical, dental, and vision coverage for employee and dependents; flexible PTO plus 10 paid holidays; 401(k) with match; 16-week parental leave for birthing parents (8 weeks for others); HSA/FSA contributions; life insurance and short/long-term disability; free daily lunch in-office; annual learning stipend.

Interview Timeline

Two to three weeks total: recruiter screen (30 minutes), technical assessment (45 minutes), onsite (3 hours), references. The systems design interview is explicitly AI-focused for the Applied AI role — expect to architect a RAG pipeline or multi-agent workflow under clinical constraints.

Technical Stack and Scaling Challenges Across 25 States

Sprinter Health's engineering challenge is making a hybrid virtual/in-home care platform operate reliably across 25 state regulatory regimes while field clinicians collect PHI on iPhones in patients' living rooms.

Device Fleet and Identity: Two-Person IT, 400+ Employees

When Richard Patterson joined as the first dedicated IT hire in August 2025, he inherited a 400-person company with no IT department. Field clinicians, known as "Sprinters," carry iPhones and iPads to perform phlebotomy and nursing visits, accessing PHI at the point of care. Corporate staff work remotely on Macs. Both fleets touch protected health information.

Patterson reports to the COO with a mandate: support aggressive growth for 18 months without proportional headcount increases. The compliance baseline is non-negotiable: annual SOC 2 Type 2 audits, continuous HIPAA compliance across every device, and a first HITRUST audit targeted for Q1 2027.

The Mac fleet migrated to Iru (an MDM platform) within months. Patterson applies CIS benchmarks automatically through Iru's Blueprint system — "the ability to go in and automatically apply CIS guidelines to a blueprint right out of the gate, to be able to show that to compliance as we're getting ready to go through audits, it's just peace of mind." Iru's Passport feature binds Mac authentication to Google Workspace: when an employee's Google account is terminated, laptop access revokes instantly without manual MDM intervention.

Device management now consumes roughly 10% of Patterson's time. His direct report spends ~50% of hers managing the iPhone/iPad fleet still on the legacy MDM — the same operational burden that drove the Mac migration. The iOS migration is planned for this year.

State-by-State Regulatory Heterogeneity

Expanding to 25 states means navigating distinct clinician licensing and scope-of-practice rules for in-home visits, telehealth prescribing and supervision requirements, patient consent frameworks for AI-augmented workflows, and data residency and breach notification timelines (some stricter than HIPAA's 60-day federal floor). Texas, for example, has enacted multiple laws impacting AI use in healthcare, including explicit consent requirements before sensitive information (mental health, reproductive health, substance use disorder, genetic data) can be processed by automated systems. The HIPAA Journal notes that for multi-state operators, workforce training must reflect the most protective state-level requirements, and warns that PHI disclosed to a third-party AI tool without a Business Associate Agreement qualifies as a notifiable breach.

Real-Time Coordination: Virtual Team ↔ In-Home Sprinter

The care model pairs a virtual clinical team with field Sprinters. A visit generates vital signs, lab orders, phlebotomy results, clinical notes, and care plan updates — all flowing from an iPhone in a patient's home to virtual physicians. The system matches visit requests to available Sprinters within geographic and licensing constraints, surfaces relevant patient history before arrival, streams findings to the virtual clinician during or immediately after the visit, and triggers follow-up actions before the Sprinter leaves the home.

Data Integration: Aggregating the Visit

Each visit touches lab integrations, EHR write-backs, claims and eligibility verification, and quality measure reporting for payer contracts. Sprinter's clinical operations writing describes a "closed-loop care pathway that includes clinical data aggregation, comprehensive in-home treatment, and care navigation." Building that loop across 25 states means normalizing lab result formats, mapping local clinician credentials to payer networks, and maintaining audit trails for every PHI touchpoint — all while the AI Automation Team ships new decision-support features.

HIPAA Architecture: Minimum Necessary, BAAs, Audit Logs

HIPAA's Security Rule is technology-neutral, but operational risks shift with AI. The HIPAA Journal outlines four AI categories in healthcare (autonomous, augmented, automation with AI, generative), each with distinct PHI exposure surfaces. Every third-party AI vendor touching PHI needs a Business Associate Agreement. The compliance automation tooling (integrated with Iru) auto-checks controls like screensaver timeouts and pulls audit evidence for SOC 2 Type 2; auditors require proof controls operated continuously, not just at review time.

Best-practice architecture for HIPAA-aware AI agents recommends: PHI isolation at the node level, immutable audit trails via state persistence (LangGraph checkpointers), human-in-the-loop checkpoints for clinical recommendations, and BAA-compliant API gateway routing (Azure OpenAI, AWS Bedrock, Vertex AI). Sprinter's AI Automation Team is building this architecture into the care platform.

The Scaling Bottleneck

Two-person IT. 25 states. 400+ employees. A field fleet on legacy MDM. A virtual care team that grows with every state launch. HITRUST preparation consuming recovered IT capacity. The AI Automation Team's mandate (logistics orchestration, clinical decision support, visit lifecycle automation) is the lever that lets the platform scale without linear headcount growth. But every new state adds licensing logic, consent variants, payer integration, and audit scope. The engineering problem isn't the model. It's the compliance architecture that lets the model run in production.

Clinical Operations Integration: Engineers Embedded with Care Teams

Sprinter Health's leadership roster reads like a joint command: VP of Engineering Jiahan Ericsson sits alongside VP of AI Scott Fleming and GM of AI Product Mitt Coats, while Chief Medical Officer Melissa Welch, Chief Product Officer Justin Larkin, VP of Clinical Strategy Steve Dalvin, and VP of Patient Operations Chelsea Grimme run the clinical and operational sides.

The hybrid visit architecture forces this integration. A Sprinter, a W-2 phlebotomist cross-trained in medical assistant and community health worker skills, arrives at a patient's home with a tablet. That tablet connects to a virtual care team of nurse practitioners, physicians, pharmacists, and care navigators. When the Sprinter draws blood, performs a 12-lead ECG, or screens for fall risk, the virtual clinician sees the data in real time and can intervene, order additional diagnostics, or adjust the care plan on the spot. The logistics AI that routed the Sprinter to that home, matching patient needs to clinician availability, travel time, and state licensure, runs on the same platform that surfaces the patient's aggregated HIE, claims, lab, and pharmacy history before the visit starts.

This closed loop is where the AI automation hires land. The Series B capital targets engineers who can build the connective tissue between the Sprinter's tablet, the virtual clinician's dashboard, and the care navigator's follow-up queue. The 80% gap closure rate and 90+ NPS Welch reports are the output of this engine.

Expansion Roadmap: From Bay Area Pilot to Nationwide Network

Sprinter Health launched Care+ visits in California in 2024. By February 2026, the model reached nearly 25 states, roughly half the country, the company announced. Expansion moved in two dimensions at once: geography and lines of business. Medicare, Medicaid, and Marketplace plans all came online during the same period, broadening the addressable member base beyond the initial commercial focus.

The supply chain for in-home visits runs on a W-2 workforce, not contractors. "Sprinters" are licensed phlebotomists hired from the communities they serve, then trained in those skills. Each carries a tablet connecting them to virtual nurse practitioners during visits. Logistics AI routes them using route simulations that match patient demand to qualified, available Sprinters in the same area. The company reports 98% on-time arrival rates across the network.

New states launching in 2026 will follow the same pattern. Dr. Melissa Welch, CMO, has said only that "more launching this year" — a signal that the infrastructure built in the Bay Area is now being stress-tested at national scale.

Why AI Automation Is the Differentiator

The home-care market splits into tiers. Traditional agencies staff visits with contractors and paper schedules; they don't build software. The middle tier digitized the dispatch layer: Honor partners with agencies to fill shifts, IntelyCare manages nursing rosters, Clipboard Health runs a nurse marketplace. DispatchHealth, at $733 million raised, sends clinicians for same-day urgent visits but stops at the door; its co-pay model ($0–$45) targets episodic acuity, not longitudinal gap closure.

Sprinter sits in a narrower band: full-stack medical practice plus platform. Its "Sprinters" are W-2 phlebotomists sourced from the communities they serve, trained in both medical-assistant and community-health-worker skills. The company reports a 30 percent member booking rate. Those numbers only hold if the logistics engine (routing, eligibility verification, kit prep, clinical documentation) runs without human triage at every step.

That engine is what the AI automation team builds. Cityblock Health ($891 million raised) and ConcertoCare ($150 million) employ care teams but lean on external EHRs and manual care-plan workflows. Sensi.AI ($93.7 million) sells an AI virtual agent to agencies for 24/7 monitoring; it doesn't employ clinicians or own the visit. PathAI and Tempus excel at diagnostic AI but don't route a phlebotomist to a living room. Sprinter's differentiation is the closed loop: the same system that schedules the visit also prompts the retinal exam, flags the missing colonoscopy, and feeds the result back to the health plan.

The hiring wave reflects that scope. A staff engineer on the AI automation team owns logistics orchestration across 25 state regulatory regimes; a senior engineer builds clinical decision support that meets HIPAA and FDA boundaries. DispatchHealth hires clinicians and dispatchers. Honor hires marketplace operators. Sprinter hires engineers who sit beside nurses and pharmacists (Mitt Coats as GM of AI Product, Scott Fleming as VP of AI, Jiahan Ericsson as VP of Engineering) because the product is the care delivery itself.

The moat isn't the visit. It's the system that makes the visit repeatable, auditable, and profitable across every zip code.

Working in frontier tech? Zero G Talent tracks the openings: browse frontier tech jobs, openings at Sprinter Health, and the people building the field.