POC Lab Portal
Portal contains Twins, actuall Stack runs in lab environment
A unified Lakehouse Architecture that converges health data streams from multiple portals and source systems into a central data lake. Raw ingested data is progressively refined through a Medallion design — Bronze, Silver, and Gold layers — ensuring data quality, traceability, and readiness for analytics. The Gold layer feeds a real-time visualization layer delivering actionable insights via an interactive digital twin dashboard.
A world-class Hospital Digital Experience built at the intersection of clinical expertise and academic research excellence. Leveraging cloud-native architectures and resilient infrastructure design, this platform unifies patient journeys, operational workflows, and clinical intelligence into a single, adaptive digital fabric — meeting the demands of both frontline care and cutting-edge medical research. A Digital Twin has been created to emulate this experience, allowing stakeholders to explore and validate every layer of the hospital's digital ecosystem in a safe, simulated environment.
A practical network resiliency testbed for enterprise DC/DR and BCP programs. Critical network elements — workstation, DNS, L2 switch, L3 gateway, ILB, and BCP app segment — are virtualized and simulated using Linux namespaces to mirror production-style segmentation without physical hardware dependencies. Teams can validate failover readiness through DNS re-pointing and firewall policy-based routing (PBR), then observe route transitions and service continuity in an animated twin before applying strategies in live environments.
Using data talker, Public datasets can be pulled in, processed through ETL pipelines, and shaped into analytics-ready snapshots. From there, a Data Talker can be built using PandasAI, Ollama-hosted local LLM models, and smart AI routers powered by compact SLM decision layers to answer questions, generate charts, and guide users through curated analytical flows. Because the models run locally, there is no external token consumption cost, and if demand grows, external GPU acceleration can be added to scale inference performance for richer workloads.
A lightweight, modular edge telemetry platform capturing sensor data from distributed devices and consolidating streams into unified, real-time insights. Built with WebSocket-based ingestion and FastAPI async architecture, the platform processes device motion, orientation, and environmental metrics with sub-100ms latency. Designed for extensibility — integrate MQTT brokers, ML prediction pipelines, and alert generators as your solution scales.