.png&w=256&q=75)
Jhair Alexis Chapilliquen Ruiz@Alex-ChapR
1
Events attended
1
Submissions made
Peru
1 year of experience
About me
Me apasiona construir software que resuelva problemas reales y que pueda mantenerse, escalar y evolucionar con el tiempo. A lo largo de mi experiencia he participado en el desarrollo de plataformas empresariales, soluciones SaaS, sistemas de monitoreo industrial, integraciones ITSM, arquitecturas basadas en APIs y proyectos de observabilidad. He trabajado desde la implementación de interfaces y servicios backend hasta despliegues, monitoreo, infraestructura e integración con entornos productivos. Actualmente trabajo principalmente con Python (FastAPI y Flask), React, SQL Server y tecnologías de observabilidad como OpenTelemetry, Prometheus, Loki, Tempo y Grafana. Disfruto involucrarme en los problemas de principio a fin, entender el contexto del negocio y encontrar soluciones que aporten valor real. También me interesa acompañar a otros desarrolladores, compartir conocimiento y contribuir a entornos colaborativos donde la calidad del software sea una prioridad. Interesado en arquitectura de software, sistemas distribuidos, observabilidad e inteligencia artificial aplicada. Busco seguir creciendo hacia posiciones de liderazgo técnico donde pueda contribuir tanto en la toma de decisiones arquitectónicas como en el desarrollo de soluciones de alto impacto.
🤝 Top Collaborators
.png&w=640&q=75)
🤓 Latest Submissions
AXONFAS: Clinic-Supervised Home Physiotherapy
Patients who do their physiotherapy at home do it alone: nobody checks their form, unsafe compensations go unnoticed, and the clinic only finds out at the next visit. AXONFAS brings the clinic into the living room. What works today (web app in English and Spanish, synthetic data): - Red-flag triage before self-guided camera exercise: one red flag blocks the routine and points to emergency care. - Camera setup: checks framing, distance and angle live and turns off measures that angle can't see reliably. - Monitored sit-to-stand: 3 calibration reps set the patient's baseline, then a deterministic rules engine flags trunk flexion and knee valgus in real time, logs each alert's latency and draws the correction on a mirrored skeleton. - 24 exercises from public sources (AAOS, NHS, CDC STEADI) with animated 3D muscle figures, a watch-and-follow tutorial, rep counting and a no-camera routine. - Physiotherapist and patient accounts: invite code with consent, a plan with safety limits, today's plan, a 0-10 pain score and a live view with skeleton and muscle map. A pain button stops the session at once. - SOAP note drafts built only from stored session data, signed by the physiotherapist and exported as FHIR R4. Clinic accounts with professional licenses. - Guest access without sign-up. Video never leaves home: only 2D keypoints are sent. AI on AMD: every LLM call goes through the AMD Token Factory API (open models, with automatic fallback). The clinical agent proposes plan changes as structured JSON, and the physiotherapist gets an AI progress summary. Alerts come only from the rules engine, never from the LLM, and no AI change can leave the safety limits the physiotherapist set. Deploy: Docker with automatic HTTPS on AMD Developer Cloud. Next: the whole clinic on one AMD Instinct MI300X with ROCm (vLLM, Whisper, TTS, 3D pose); that stack is scripted and ready to switch on. Staging only. Not a medical device.
18 Oct 2026
