What SaMD classes do you work on?

We engineer software across SaMD classes, with the most experience on Class IIa / Class II equivalents. We adapt the lifecycle depth to your classification and intended use.

How do you implement IEC 62304?

We tailor the lifecycle to your software safety class — planning, requirements, architecture, unit/integration/system verification, problem resolution, configuration and release — with traceability throughout.

How do you handle PHI and clinical data?

We design for least-privilege access, encryption in transit and at rest, audited data flows, and de-identification where possible. We work under your data-processing agreements and security baseline.

Can models be retrained after release?

Yes — with a defined change-control and revalidation strategy. We help you choose between locked, periodically retrained, or continuous-learning pathways and document the rationale.

On-prem, cloud, or hybrid?

All three. We pick the topology that fits your clinical, regulatory, and operational constraints, and engineer accordingly.

§ 02 — Services02 / 04

Two practices.
One standard.

We help companies ship AI features and regulated medical software with the same engineering rigor — so your product is built for clinic from day one.

CHAPTER 01 — AI

AI consulting & development

Pragmatic AI engineering — from first prototype to a model your team can monitor, retrain, and trust in production.

01AI strategy, discovery and feasibility studies
02Custom machine learning and deep learning models
03LLM systems: RAG, agents, fine-tuning, evaluation
04Computer vision for clinical imaging and signals
05Data engineering, labeling pipelines, and MLOps
06Model monitoring, drift detection, and lifecycle ops

CHAPTER 02 — SaMD

SaMD development

Software as a Medical Device, engineered with the lifecycle, risk management, and traceability your auditors and clinics expect.

01IEC 62304 software lifecycle implementation
02ISO 14971 risk management and traceability
03FDA and EU MDR documentation support
04Software architecture for safety-critical systems
05Verification, validation, and clinical evaluation support
06Cloud, security, and quality management integration

§ 03Engagement models

How we work with you.

Consulting

Short, focused engagements: strategy, feasibility, technical audits, regulatory pathway.

Custom build

We design and deliver your AI or SaMD product end-to-end, from prototype to production.

Embedded team

Senior engineers integrated into your team, with our process and rigor on top of yours.

§ 04Industries

Sectors we serve.

RADRadiologyImaging triage, segmentation, reporting copilots.
PATPathologyWhole-slide imaging models and workflow integration.
CARCardiologySignal processing, arrhythmia detection, monitoring.
GENGenomicsVariant pipelines and clinical reporting.
LABLab automationInstrument software, data pipelines, QC analytics.
PHXPharmaTrial analytics, RWE pipelines, regulatory tooling.
DTXDigital therapeuticsPatient-facing SaMD with adherence and outcomes.
ENTEnterprise AIInternal copilots and ML platforms for regulated orgs.

§ 05Deliverables

What you receive.

Every engagement ships engineering artifacts that survive audits, scale, and team turnover.

#TypeArtifactStandardIncl.

01Architecture & designInternal

02Risk management file inputsISO 14971

03Software requirements & traceabilityIEC 62304

04Model cards & eval reportsInternal / FDA GMLP

05Verification & validation evidenceIEC 62304 §5.6–5.7

06Deployment runbooks & SBOMsInternal / IEC 81001-5-1

07Cybersecurity documentationFDA / EU MDR

08Release notes & configuration itemsIEC 62304 §8

§ 06Tech stack

Tools we reach for.

§ 07FAQ

Good to know.

What SaMD classes do you work on?: We engineer software across SaMD classes, with the most experience on Class IIa / Class II equivalents. We adapt the lifecycle depth to your classification and intended use.
How do you implement IEC 62304?: We tailor the lifecycle to your software safety class — planning, requirements, architecture, unit/integration/system verification, problem resolution, configuration and release — with traceability throughout.
How do you handle PHI and clinical data?: We design for least-privilege access, encryption in transit and at rest, audited data flows, and de-identification where possible. We work under your data-processing agreements and security baseline.
Can models be retrained after release?: Yes — with a defined change-control and revalidation strategy. We help you choose between locked, periodically retrained, or continuous-learning pathways and document the rationale.
On-prem, cloud, or hybrid?: All three. We pick the topology that fits your clinical, regulatory, and operational constraints, and engineer accordingly.

§ 08Next step

Ready to scope your project?

Talk to an engineer

Two practices.One standard.