01Production
Medicus.IA
A health operating system.
Longitudinal records, medication safety and a deterministic clinical engine with FHIR R4 interoperability.
21 modules · 153 clinical routes
MEDICINE + ENGINEERING + JUDGEMENT
Intelligence
that can
answer for itself.
I’m Dr. Ixequi Luna. Physician and AI architect. I turn critical workflows into traceable software, from clinical decisions to business operations.
Physician · CEO & CTO at Firmus.IAPUEBLA, MEXICO · GLOBAL COLLABORATION
+ HUMAN JUDGEMENT01
RULES · TRACEABILITY · EVIDENCE
FROM DOMAIN TO SYSTEM / IX.202613products built
90+AI engines designed
3,208automated tests
61learning credentials
Figures reported in my 2026 CV; they span different projects and stages, and are not live measurements.
01 / SELECTED WORK
From a real problem
to a system that delivers.
Healthcare, business operations and interoperability. Architecture and code written by me.
No projects match that search.
02Production
PuestoYa
Money demands exact rules.
HR and payroll for a four-company group. Labour calculations, frozen billing periods and an immutable audit log.
232 tests · 95.1% coverage
05Public code
health-interop-lab
Interoperability you can run.
One clinical workflow across five languages: HL7 v2, Kafka, FHIR R4, gRPC and EDI X12, instrumented with OpenTelemetry.
71 controls as tests¹
More systems. The same rigour.
Explore the scope of each project.
Viajare AI Travel OSPartial production
From booking to reconciliation.
Booking audits, customer service, CRM and commercial operations. Engine 01 in production and three Engine 02 surfaces live.
1,609 tests · 4 engines designed
EmpoweringBizDelivered
32 engines. One strategy.
Orchestration with typed signals, boundary validation and one synthesis authority. Under support and maintenance.
956 documented tests
Pakal AI CRMDelivered
AI proposes. People decide.
Multi-company executive CRM integrated with Odoo: quotes, receivables, multi-currency reconciliation and WhatsApp orders.
5 modules delivered
PendientIAProduction
One shared operation.
Tasks, boards, invitations and reports. Five roles with workspace isolation and activity auditing.
5 roles · multi-company isolation
Caro de NardoIn build
Commerce with an identity.
A digital experience for handcrafted Tuscan perfumery: a bespoke storefront and Medusa v2 commerce engine.
Creative and technical direction
8SanoIn development
Wellbeing for everyday life.
Habit plans, progress tracking and e-commerce in one experience.
Product in development
AgricusIA · JuricusIAUpcoming
Domain-specific intelligence.
Predictive AI for agriculture and contract analysis. Products in development, with staged scope and validation.
Agriculture / Legal
EstudicusIA · EdifikusIA · Lattixe360GUpcoming
New industries. The same rigour.
Adaptive learning, real-estate analytics and marketing automation within the Firmus.IA portfolio.
Education / Real estate / Marketing
Status and metrics from the 2026 CV. ¹ SOC 2 controls expressed as tests; they do not constitute SOC 2 certification or an audit report.
Also built
Grace, Viñedos El Sueño, IA Mundial and the Firmus.IA and Medicus.IA web experiences. The CV covers the full portfolio.
02 / HOW I THINK
Trust is designed.
Then put to the test.
Judgement before automation.
I start with the domain: who decides, what evidence they need and what harm an error could cause.
Rules where reproducibility matters.
I separate language tasks from critical decisions. An output should be traceable to its rule and source.
The ability to stop.
Incomplete data and missing context should lead to explicit abstention, not a persuasive answer.
Security throughout the lifecycle.
Data contracts, least privilege, regression tests, traceability and reversible deployments.
CASE / MEDICUS.IA
More time for
the person.
In Pilot 0, as reported in the CV, documentation time fell from 60 to 15 minutes per patient across 100 consultations. This is a pilot result, not a guarantee of clinical outcomes.
Before60 min
With Medicus.IA15 min
−75 % documentation time
Source: 2026 CV · Pilot 0 · n = 100 consultations
03 / THE PERSON BEHIND THE CODE
A physician by training.
A builder by conviction.
Clinical practice taught me to listen, recognise uncertainty and take responsibility for a decision. Those same principles guide my engineering.
I hold a medical degree from BUAP, lead Firmus.IA as CEO and CTO, and architect Medicus.IA. I maintain a clinical practice and write the core of my systems in Python, TypeScript and Go.
Ixequi LunaMD · CEO & CTO · AI ARCHITECT
- 2026
Firmus.IA · CEO & CTO
Technology leadership, architecture and sovereign AI development for critical industries.
- 2026
Client systems architecture
PuestoYa, Viajare, Pakal and EmpoweringBiz; product, architecture and security review for Yaan Technology.
- 2025 →
Medicine and research
Private and occupational practice in Val’Quirico. Co-founding an ambulatory surgery network and leading research.
- 2016–2024
Faculty of Medicine · BUAP
Medical degree. Internship at Hospital General del Sur and research service at the Biomolecular Detection Center.
Engineering
Python · TypeScript · Go · Java · Kotlin · C#
Interoperability
FHIR R4 · HL7 v2 · gRPC · Kafka · EDI X12
Platforms
FastAPI · Next.js · PostgreSQL · AWS · Docker
Quality & security
CI/CD · Playwright · OpenTelemetry · RBAC · SHA-256
04 / RESEARCH & LEARNING
Curiosity,
with a method.
Scientific evidence and continuous learning are part of the way I build.
Autism Spectrum Disorder and Temporal Lobe Epilepsy in a Physician: A Self-Reported Case of Neuropsychiatric and Functional Adaptations.
Luna-Mendoza I, Torres-Rasgado E, Cerón-Morales JA.
DOI 10.7759/cureus.93350
PMID 41158905 · PMCID PMC12558430
CONTINUOUS LEARNING
61 credentials.
A practice that evolves.
- IBM · AWS · Anthropic · Google
- Generative AI, cloud, data and user experience.
- Johns Hopkins · Yale · KAIST
- Neuroscience, medical software and imaging.
- Glasgow · Rice · Northeastern
- Biomedical visualisation, terminology and healthcare management.
05 / WORK TOGETHER
From a difficult question
to the next step.
Architecture & engineering
From discovery to system: domain, contracts, implementation, validation and operations.
Discuss scopeTechnical due diligence
Product, security, data integrity and delivery reviews. Prioritised findings and a remediation path.
Discuss scopeClinical AI & interoperability
Decision support, documentation and FHIR/HL7 workflows, with professional oversight and use-case validation.
Discuss scopeTechnology leadership
Fractional CTO, product strategy, mentoring and engineering governance for teams ready to move forward.
Discuss scopeReference fees & engagement options
Rates previously published on this site. Scope, availability, taxes and terms are confirmed in a proposal before engagement.
Advisory · 1-hour minimum$150 USD/h
Engineering · 2-hour minimum$200 USD/h
AI assessment · 1–2 weeks$2,500 USD
Prototype · 3–6 weeks$8,000–15,000 USD
Platform · 2–6 months$25,000–75,000 USD
Transformation · 6–12 months$75,000–150,000+ USD
Retainer · 15 / 30 / 60 h monthly$3,000 / $6,000 / $12,000 USD
Initial call: 30 minutes, no charge. Previous project payment references: 40% upfront, 30% at milestone and 30% at delivery; retainers monthly in advance. Healthcare projects require a scoped validation and regulatory budget.
The earlier AI subscription plans have no verified signup and payment service on this site. Ask about availability before subscribing.
Feedback published on the earlier site
In love with the intelligence, delivery speed, and quality of the AI architecture he built for my company.
A genius. It's incredible how easily he programs and delivers results better than the biggest AIs on the market.
I've been at MIT working with top-tier AI engineers. Ixequi is 10x more efficient, skilled, and innovative.
His products are gold. Literally gold in code. His AIs are complex, incredibly complete, and truly revolutionary.
Ixequi is a machine. He's beyond my level and speed. He revolutionized my family business with his mentoring and consulting.
I've been coding for 20 years and contributed to the WHO. He surpassed me in just one week. His approach to coding is revolutionary.
A STARTING POINT
Give your
project shape.
Prepare a brief for our first conversation. It is organised in your browser: no AI, no invented return on investment and no automatic submission.
Describe the process without including patient data, passwords or confidential information.
Your starting point
Opening email only prepares a draft. You decide whether to send it.
06 / THINKING IN PUBLIC
Notes from
the build.
01 / 04.2026
Deterministic AI and clinical decisions
02 / 03.2026Data sovereignty and local inference
03 / 03.2026How to orchestrate 32 AI engines
Archive articles, March–April 2026. Their figures describe that stage; the CV covers subsequent progress.
BEFORE WE TALK
Common questions.
Do you work with international teams?
Yes. Based in Puebla, Mexico, I collaborate remotely or in a hybrid arrangement. I work in Spanish and English; scope and time zones are agreed upfront.
Does everything need AI?
No. I evaluate the problem first. A clear rule, a better integration or a simpler workflow may be the better solution.
Does AI replace medical judgement?
No. In my approach, the system suggests and documents; the professional retains judgement and signs the decision. Validation depends on the use case.
How does an engagement begin?
With a 30-minute conversation about the process, available evidence and desired outcome. We then define scope, deliverables, risks and budget.
THE NEXT SYSTEM STARTS WITH A CONVERSATION
Let’s build something
that matters.
Architecture, technology leadership or a second look at a critical system. Tell me what you need to solve.
Scheduling opens on Calendly, an external service.