dr Ferdi Iskandar · https://ferdiiskandar.com |
Founder, Sentra Artificial Intelligence
|
I'm Codieverse. By training I'm a physician; by circumstance I ended up running a hospital, and along the way started building Sentra Artificial Intelligence, an attempt at a healthcare AI ecosystem covering clinical reasoning, medical workflow automation, patient journey intelligence, and Indonesian healthcare infrastructure. Most of what I do lives at the overlap of clinical judgment, AI infrastructure, and the messier realities of hospital operations.
My work sits at the intersection of clinical judgment, AI infrastructure, and real-world hospital operations.
|
Medical reasoning, triage, diagnosis support, guideline-aware workflows, and safety gates for clinician-facing systems. |
RAG, memory, orchestration, voice agents, autonomous workflows, observability, and modular healthcare AI engines. |
Hospital dashboards, EMR bridges, coding audit, admission flow, bed management, telemedicine, and real-world deployment discipline. |
The mission is simple but demanding: turn fragmented healthcare workflows into intelligent, auditable, assistive systems that clinicians can trust.
AI in medicine should not be theatrical. It should be useful, humble, explainable, and safe.
|
AI supports the doctor. It does not claim final authority. Every serious medical AI system must preserve the human reviewer as the final accountable decision-maker. |
Clinical output must pass deterministic checks, red-flag detection, contraindication review, uncertainty handling, and escalation boundaries. |
|
Indonesian healthcare realities matter: primary care, BPJS complexity, EMR friction, referral pathways, hospital operations, and patient journey fragmentation. |
Diagnosis, RAG, memory, EMR automation, telemedicine, cybersecurity, and dashboards must remain separable, auditable, and replaceable. |
|
No overbuilt fantasy platform. Start with a real clinical or operational pain, solve it clearly, then expand deliberately. |
Clinical AI must show its inputs, reasoning boundaries, confidence, uncertainty, evidence trail, and escalation logic. |
|
Autonomous Artificial Diagnostic Intelligence A layered autonomous diagnostic reasoning engine designed to assist clinical decision-making through structured differential reasoning, safety checks, ICD mapping, and clinician-review boundaries. Core intent: diagnostic support, not final diagnosis. |
Voice-First Clinical Intelligence A real-time voice clinical assistant built to support medical encounters, capture clinical context, surface structured insights, and reduce documentation friction during consultations. Core intent: ambient clinical assistance. |
|
Unified Clinical Operations Platform A central operations dashboard for EMR workflows, ICD support, reporting, communication, patient monitoring, telemedicine, and clinical intelligence surfaces. Core intent: one command center for clinical operations. |
Clinical Workflow Automation Extension A browser extension for automating clinical workflow steps, transferring structured data into EMR systems, and assisting clinicians with decision-support surfaces inside their existing workflow. Core intent: reduce repetitive EMR friction. |
|
Remote Consultation Infrastructure A WebRTC-based remote consultation system with clinical notes, scheduling integration, patient-room flow, and consultation continuity. Core intent: make distance care clinically usable, not merely video-based. |
Referral & Awareness-Intelligence Protocol A referral-routing and claim-awareness system for healthcare operations, designed around regulatory fluctuation, insurance/BPJS complexity, semantic cache support, and contextual decision assistance. Current engineering note: scaffold exists, but requires final verification, refactor decision, and environment/security audit before being called production-ready. |
|
Persistent Clinical Memory Layer A memory architecture for AI agents that preserves useful clinical context across sessions using semantic retrieval, persistent memory, and structured recall. Core intent: make AI agents context-aware without unsafe hidden assumptions. |
Maternal & Pediatric Personal Virtual Agent A personal virtual agent for patient accompaniment from pre-conception through pregnancy, postpartum care, and pediatric continuity. Core intent: proactive patient guidance across the maternal-child journey. |
|
Hospital Interoperability Dashboard A cross-unit interoperability dashboard for RSIA Melinda operations, designed to unify fragmented departmental data and make operational status visible. Core intent: reduce hospital data silos. |
Predictive Cybersecurity Architecture A layered cybersecurity architecture for protecting clinical data through monitoring, encryption, behavioral analysis, access governance, and rapid containment concepts. Core intent: secure the clinical AI operating environment. |
|
Admission & Patient Journey Automation An automated admission system for extracting referral documents, validating schedules, tracking patient journey steps, and reducing queue friction. Core intent: move admission from manual queue to intelligent flow. |
Structured Early Risk Detection An early triage layer that collects structured symptoms, flags emergency risk, and prepares the clinician before face-to-face consultation. Core intent: detect risk earlier and prioritize care. |
|
Post-Discharge Continuity Engine A proactive monitoring system for post-discharge follow-up, complication prevention, immunization reminders, and continuity-of-care workflows. Core intent: prevent patients from disappearing after discharge. |
Voice-to-EMR Documentation Engine A clinical documentation engine for converting consultation speech into structured medical notes and mapping relevant information into EMR fields. Core intent: let doctors focus on patients, not typing. |
|
Early Warning & Escalation Layer A clinical alert system that watches laboratory data, telemetry, and high-risk signals to support rapid escalation to the right clinical team. Core intent: prevent critical deterioration from being missed. |
Discharge & Bed Readiness Orchestration An orchestration layer for discharge readiness, bed turnover, housekeeping, pharmacy, billing, and department coordination. Core intent: improve patient flow and bed availability. |
|
Clinical Coding & Claim Defense An automated audit system for checking clinical coding consistency against documentation, reducing claim dispute risk and improving coding reliability. Core intent: defend claims with cleaner clinical evidence. |
Operating Room Logistics Intelligence A real-time operating room orchestration system for priority cases, operating team readiness, blood product logistics, room allocation, and case flow. Core intent: coordinate high-stakes surgical logistics. |
|
Pregnancy Observation & Guidance System A pregnancy observation system for maternal-fetal monitoring, risk detection, structured antenatal context, and escalation support. Core intent: safer maternal-fetal surveillance. |
Clinical Decision Orchestration System A clinical decision orchestration engine that maps guidelines into diagnostic and operational workflows while preserving clinician oversight. Core intent: turn guidelines into usable clinical flow. |
|
Algorithmic Emergency Prioritization A severity-score-based triage layer for emergency prioritization, risk stratification, and structured routing. Core intent: prioritize patients by risk, not by noise. |
Clinical Risk Forecasting Engine A predictive engine for estimating deterioration, complications, readmission risk, and clinical trajectory changes. Core intent: shift care from reactive to anticipatory. |
|
|
|
CLINICAL INPUT
│
▼
STRUCTURED CAPTURE
│ complaints · vitals · labs · history · documents · voice
▼
NORMALIZATION LAYER
│ terminology · units · ICD · FHIR-aware structures
▼
REASONING / RAG / MEMORY
│ clinical engine · retrieval evidence · persistent context
▼
SAFETY GATE
│ red flags · uncertainty · contraindication · escalation boundary
▼
CLINICIAN-FACING OUTPUT
│ summary · differential support · triage signal · workflow action
▼
HUMAN REVIEW
The architecture is intentionally conservative: AI proposes, structures, retrieves, and assists. Clinicians decide.
|
React, Next.js, Tailwind, clinical dashboard UX, calm enterprise interface, strong hierarchy, readable data surfaces, and low-friction workflows. |
TypeScript-first services, modular APIs, auditable contracts, clean package boundaries, and explicit integration layers. |
|
RAG, orchestration, agent memory, local-first components where possible, model-agnostic architecture, and safety-aware output formatting. |
Real-time clinical voice capture, voice-to-EMR workflows, OCR/document ingestion, structured clinical note generation, and review-first automation. |
|
Structured contracts, retrieval-ready documents, PHI/PII boundaries, least privilege, audit trail, and no unsafe logging. |
Prototype-to-production separation, verification before promotion, rollback awareness, and operational realism. |
No magic without audit.
No diagnosis without clinician review.
No workflow automation without rollback.
No clinical data without security boundaries.
No platform expansion without one clear problem first.
| Question | Required answer |
|---|---|
| What clinical problem does this solve? | A specific workflow pain, not a vague AI ambition. |
| Who is the human reviewer? | Doctor, nurse, admin, verifier, or operator must be clear. |
| What is outside the scope? | Non-scope prevents unsafe expansion. |
| What can fail? | Failure modes must be visible before deployment. |
| How is it verified? | Build, typecheck, test, audit, and clinical review where needed. |
|
The core engineering ecosystem for Sentra Artificial Intelligence: shared packages, healthcare apps, RAG engines, governance surfaces, orchestration tools, and clinical infrastructure experiments. |
A clinical dashboard and operational command surface for CDSS, telemedicine, EMR bridge workflows, trajectory analytics, and clinical reporting. |
|
A clinical browser-assistance surface for emergency detection, diagnosis support, structured workflow automation, and side-panel clinical intelligence. |
A healthcare referral and routing surface with diagnosis endpoint concepts, semantic cache support, and memory-service helpers. |
RSIA Melinda DHAI Strategic healthcare collaboration and institutional support in the development of applied clinical intelligence systems, operational interoperability, and hospital-centered AI workflows. |
MedLab Supporting ecosystem for medical and diagnostic innovation, aligned with healthcare workflow modernization, data intelligence, and practical AI deployment. |
Dedicated to Aldebaran, Aimee, Audrey, and Del — & Indonesia Healthcare Ecosystem.
Sentra Artificial Intelligence · Healthcare AI for safer, clearer, more humane clinical systems.