Matthew Emmanuel Labrador@matthewtlabrador86476
2
Events attended
1
Submissions made
Philippines
1 year of experience
About me
Computer Science student at University of the Philippines Manila and Department of Science and Technology Scholar. Passionate about machine learning, I hold multiple certifications in prompt engineering and programming languages (including Python and C). Alongside my technical background in local LLM deployment and data preprocessing, I bring strong bilingual communication skills as a native Filipino and fluent English speaker. Always looking to tackle new challenges and connect with other builders!
š¤ Top Collaborators
.png&w=640&q=75)
š¤ Latest Submissions
Glycoswarm AI
72% of diabetics are only diagnosed after a complication has already appeared, and damage to the kidneys, nerves, eyes, or heart can develop silently for 4-7 years before that. Today's screening checks one specialist, one chart, one metric at a time. GlycoSwarm AI fixes that with a 4-agent specialist pipeline: dedicated Renal, Neuropathy, Retinal, and Cardiovascular agents each reason over a different slice of a patient's real lab panel, then a Synthesis Agent weighs all four findings against each other and outputs one ranked referral recommendation, all generated live, not pre-computed or hardcoded. Every specialist writes and executes its own Python risk-scoring code against the patient's actual labs in real time, visible in a live reasoning terminal. The system runs on genuine on-GPU inference: Gemma 4 26B served via Ollama on an AMD Instinct MI300X (AMD Developer Cloud), with Fireworks GLM 5.2 as a real, user-triggered fallback. If neither provider is reachable, it honestly reports "unavailable" rather than faking a clean result. We validated the full pipeline across 104 real, de-identified CDC NHANES patients, with 90 flagged in at least one organ system. The dashboard (Next.js, FastAPI, LangGraph) renders an anatomical risk map, live agent terminal, and an exportable clinical brief. It's a prototype, not a certified clinical device, but the architecture generalizes beyond diabetes to chronic kidney disease, metabolic syndrome, and post-transplant care, as a SaaS subscription or EHR-integrated license, against a $1T+ global health spend.
13 Jul 2026
