Federico Pacheco, MD PhD — AI operations for healthcare-adjacent operators

The leverage your team
is missing.

I find where AI compounds in your operation, ship the systems that capture it, and train your team to extend them — without touching clinical workflows or patient data.

Federico Pacheco, MD, PhD summa cum laude. A practising physician now full-time in AI operations.

The boundary is precise. By design.

This is a non-clinical practice. I build AI systems for the back-office of healthcare-adjacent businesses — the operations that move money, schedule people, and manage documentation. I do not build clinical decision tools, I do not touch protected health information, and I do not integrate with electronic medical records. Where my MD matters is in the disciplined process I bring to the operation, not in any clinical claim made by the systems we ship.

In scope

Medical billing back-offices
Practice management operations
Healthcare HR & staffing
Biotech operations
Healthcare services administration
Internal knowledge management
SOP & documentation automation
Non-PHI document processing
Operator training & tooling

Out of scope

Clinical decision support
Patient-facing diagnosis tools
EMR integration
PHI processing or storage
Medical device software
HIPAA-regulated infrastructure

Engagements use a Data Processing Agreement that explicitly forbids the client from inputting PHI into the systems we build.

Diagnosed the way patients are. Methodically.

Healthcare operations fail in the same way clinical decisions fail — when somebody prescribes before they have diagnosed. A vendor sells the practice a tool; the team never adopts it; the bill keeps recurring. The discipline I bring is the one any good clinician applies to a complicated case: listen first, rank hypotheses, prescribe the smallest intervention that resolves the most, verify the outcome in numbers.

01 — Diagnose

Start with the operator.

Two weeks. A structured walkthrough with the principal operator. A written brief naming the three to five operational levers AI would compound, in priority order, with a financial projection.

02 — Prescribe

Pick the one that compounds.

One intervention. Scope, stack, timeline, and a numeric definition of success — written before the build begins.

03 — Implement

Ship inside the engagement.

Eight to twelve weeks. Weekly working sessions with the operators who will run the system. Live before the engagement ends.

04 — Transfer

Your team runs it. And extends it.

Documentation, runbooks, a written extension guide, and follow-ups at 30 and 60 days. The compounding value lives in your employees.

Three ways to work together. Priced in plain numbers.

Pricing below is the base band. Healthcare-adjacent engagements carry a 30–80% premium, reflecting the operational complexity of the audience and the documentation discipline required.

i. Operations Diagnostic

$9,500

Two weeks · fixed fee · base band

A CEO or principal-operator walkthrough, an operations map, a prioritised opportunity list with financial projection, and a 90-day roadmap. Delivered as a written brief.

— Kickoff & principal walkthrough
— Operator interviews where they matter
— Written readout (90 min)
— 100% credited toward the Sprint

Book a diagnostic call

ii. Implementation Sprint

$28k – $45k

Eight to twelve weeks · base band

The top intervention from the diagnostic, designed and shipped as a working system. Documentation, runbooks, training, and an extension guide included.

— Weekly working sessions
— Operator training built in
— Live system before the engagement ends
— 30- and 60-day follow-ups

Discuss a sprint

Most engagements

iii. Fractional AI Operator

$7k – $10k

Per month · one day a week · base band

Embedded one day a week — owning the AI roadmap, vetting new tools, training the team, shipping small interventions continuously.

— Quarterly minimum, monthly renewal after
— Embedded in your operating cadence
— Async-first, weekly office hours
— Limited capacity, by introduction

Explore a retainer

A clinician who understands your operation from the inside.

I am Federico Pacheco — a German-licensed physician who has spent the last several years building research and operational systems around AI, including a doctoral thesis defended summa cum laude in 2026. I read your operation the way I would read a complex case: gather the signal honestly, rank the hypotheses, prescribe the smallest intervention that resolves the most, verify the outcome in numbers. That discipline is what an MD brings to your back-office, even when the systems we ship never go near a patient.

Medicine MD, Germany. Practising clinician through 2026.

Doctorate PhD summa cum laude, Leibniz University Hannover. Clinical methodology (CHRO), four-year study on antipsychotic-induced weight gain.

Operation The thesis itself was built around a multi-agent AI research system. The same architecture now underwrites client engagements.

Practice Full-time AI operations consulting since 2026. Delivering remotely to US, EU, and UK operators.

The objections worth pre-empting.

Is this HIPAA-compliant?

Honest answer: the question does not apply, because nothing I build processes or stores protected health information. HIPAA governs the handling of PHI by covered entities and their business associates — my work is in the non-clinical operational layer of your business. There is no PHI to protect under HIPAA in the systems we ship. Engagements use a Data Processing Agreement that explicitly forbids the client from inputting PHI into the tools we build, which keeps the boundary clean for both of us.

What if my team accidentally enters patient information into one of your systems?

Two safeguards. First, every system we build comes with a written boundary — what it is for, what it must not be given — that your operators are trained on during the build. Second, the engagement agreement explicitly forbids PHI input, with consequences spelled out. If your scope actually requires a system that handles PHI, that is a different engagement with a different partner — and I will tell you so honestly during the diagnostic.

Why does an MD do non-clinical operations work?

Because clinical training is, structurally, a course in diagnosing complex systems under uncertainty: gather signal, rank hypotheses, prescribe the smallest effective intervention, verify the outcome. Healthcare-adjacent operations — billing, practice management, biotech ops, healthcare HR — are exactly the systems where that discipline pays. Most consultants build first and learn second. The clinician's habit is the inverse: read first, build only what the diagnosis earns.

What does the diagnostic actually produce?

A written brief. It contains a current-state map of where your operation makes money or loses time, three to five prioritised opportunities for AI leverage with financial projections, a clean scope for the highest-leverage intervention, and a 90-day implementation roadmap. The fee credits 100% toward the Sprint if you proceed.

What if the diagnostic finds that AI is not the right move?

That is a valid and valuable outcome. Knowing it before you spend six figures is worth the diagnostic fee. The brief will tell you exactly where AI would have failed and where else to look for operational leverage. I do not prescribe interventions that are not earned by the diagnosis.

What does this require of me as principal?

Three to four hours across the two-week diagnostic — a 90-minute kickoff, one structured walkthrough of how your operation actually makes money, and a 90-minute readout. The Sprint then takes a weekly working session, generally with your principal in the first half and your operators in the second.

You are based in Germany. Does that matter?

Engagements are delivered remotely, with optional on-site visits. Billing is in US dollars. Working hours overlap the US East Coast morning to mid-afternoon. European fluency in regulatory frameworks (GDPR, EU AI Act) is an asset for operators thinking about international expansion or governance gold-standards. References on request.

Two weeks. A written brief. A decision you can make on numbers.

Book a 30-minute call. We will discuss the operation, where you suspect AI compounds, and whether a paid diagnostic is the right next step. There is no obligation, and the call itself is free.