Available for 1–2 engagements
Staff-level engineer for clinical AI teams moving from prototype to production.
I help healthtech teams turn promising AI demos into deployable workflows — with the data, integrations, evidence trails, monitoring, and production engineering needed for real users.
- FHIR / HL7 integrations
- Clinical data pipelines
- Evidence-linked AI
- Production systems
Experience across
Proven at scale
Clinical systems in production. Buyer scrutiny. Real-world load.
Why now
Clinical AI demos are easy. Production is where they stall.
The hard part starts when a buyer asks:
- Can this work on our data?
- Can we trace every output back to evidence?
- What happens when data is missing, stale, or contradictory?
- How is the workflow monitored after launch?
- Who reviews edge cases before they reach users?
- Can our clinical, security, and compliance teams trust it?
That is where I help: building the data, integration, evidence, monitoring, and production foundation clinical AI needs before pilots, procurement, or real users expose the gaps.
Phare Health — production-scale clinical data infrastructure
I helped scale interoperability across hospitals, EHRs, and payer systems — building the production data pipelines and validation that clinical AI depends on.
- Epic, Cerner & payer integrations across complex environments
- 70% less onboarding effort via GPU-accelerated de-identification & NLP
- 60% fewer data errors through stronger validation & governance
- Compliance built into architecture and delivery
How I work
What makes clinical AI production-grade
The gap between an impressive demo and a system clinicians can trust is engineering discipline. These are the layers I focus on.
Evidence-linked outputs
Per-criterion verdicts with references back to source data — not black-box answers a clinician can't audit.
Deterministic vs LLM routing
Rules and lookups where correctness is non-negotiable; models only where they genuinely earn their place.
Abstention & uncertainty
Systems that know when to say “not sure” and escalate, instead of confidently guessing on patient data.
Evaluation & observability
Eval harnesses, regression checks, and monitoring so quality is measured continuously, not assumed.
Human-in-the-loop review
Workflows designed around clinician review and override — AI as leverage, accountability stays with people.
Real-world data integration
FHIR/HL7 and messy clinical data wired in with de-identification and provenance, not happy-path demo inputs.
In their words
Trusted by engineering leaders and healthcare operators
Teams value senior technical judgment, clear communication, and the ability to turn ambiguous production problems into shipped systems.
Teddy was a fantastic member of our team. He is kind, collaborative, and knows a ton about healthcare data. Strongly recommend!
Teddy served as the senior technical force for the team and was instrumental in coaching and mentoring other members, while bringing us through challenges on an aggressive schedule.
Teddy's ability to decipher a problem and ask the right questions is an invaluable asset. His communication, technical skills, and leadership make him a positive addition to any team.
Work with me
How to work with me
Pick the level of involvement that matches your timeline, scope, and internal capacity.
Fit
You should reach out if:
Strong fit if
- You have a clinical AI prototype that needs to become pilot-ready.
- A buyer, hospital, payer, pharma, or enterprise customer is asking harder technical questions.
- Your workflow depends on messy clinical data, FHIR/HL7, EHR exports, or internal data pipelines.
- You need senior engineering capacity to own a production workstream without months of ramp-up.
- You need the system to be more defensible before real users depend on it.
Probably not a fit if
- You need a low-cost frontend-only developer.
- You are looking for generic website or marketing-site work.
- You want a large agency team.
- You need someone full-time onsite.
- You are not ready to give enough context for meaningful technical ownership.
Sensitive data handled carefully
I'm comfortable working in healthcare and regulated environments. For discovery, I prefer synthetic, anonymized, or de-identified data. For deeper work, I can align with your existing security, compliance, access-control, and BAA processes.