AI & infrastructure for population health

The systems that move goods to millions can carry care to the communities health systems reach last.

I'm Kapil Poreddy — an engineering leader working at the intersection of large-scale digital infrastructure and public health. I design AI-driven, cloud-native systems that turn predictive insight into real-world care, and I publish, mentor, and build in the open so others can take the work further.

Retail × Health
Where the work sits
Detection
Telemetry & disease, same craft
Open by default
Papers · code · mentorship
01

Open problems — build with me

These are live questions I'm actively working on. Each one needs collaborators: clinical partners, co-authors, students, or engineers who want to ship. If any of them is your world too, reach out.

Ambient early detection

Can everyday signals flag disease before symptoms?

Household retail and behavioral data as an early, privacy-preserving signal for conditions like chronic kidney disease and developmental delay — catching risk years before clinic-based screening does.

Collaborate on this →
Rx-by-Drone — last-mile medication access

Can aerial delivery infrastructure close the gap in medication access?

My thesis: the last-mile aerial delivery infrastructure now proven for retail can be re-architected to carry medications, diagnostics, and time-critical supplies to rural and underserved communities — designed around the constraints healthcare demands. I'm developing this as an independent research direction and inviting collaborators (health systems, researchers, logistics partners). Angles I'm exploring: privacy-preserving routing; cold-chain & chain-of-custody for regulated payloads; equity by design (reaching underserved areas, not just easy zip codes); safety as a design constraint for BVLOS-era reliability; measuring outcomes (medication adherence, access closed) rather than deliveries.

An independent research perspective, built from public-health evidence and publicly reported industry developments.

Safe AI at scale

How do autonomous agents stay in bounds?

Guardrails, golden paths, blast-radius limits, and observability that let AI agents operate safely when they're touching systems that millions of people depend on.

Collaborate on this →
02

Research highlights

Selected peer-reviewed and preprint work on AI for health and retail infrastructure. Full list on Google Scholar.

Health AI · Pilot study

AI-Enabled Early Detection of Chronic Kidney Disease in Underserved Communities Using Social Determinants of Health

Why it matters: 37M U.S. adults have CKD and most don't know it, with 2–3× higher rates in underserved communities.

Read on Scholar →
Health AI · Digital phenotyping

RetailHealth: Privacy-Preserving Digital Phenotyping for Early Developmental-Risk Estimation

Why it matters: One in six children shows developmental delay, most caught later than optimal. Uses household purchasing patterns as an ambient behavioral signal.

Read on Scholar →
Systems · Reliability ML

Reducing MTTR and Alert Fatigue with PyTorch-Powered Anomaly Detection

Why it matters: The same detection craft, pointed at infrastructure. Presented at PyTorch Conference 2025 (Linux Foundation).

Verified badge →
03

Talks & webinars

Selected sessions at IEEE and industry venues. Full speaker profile on Sessionize.

Mar 2026 · IEEE TEMS

How to Use AI Tools

IEEE TEMS "HOW TO" Series, Episode 10 — practical AI for productivity, development, and decision-making.

Oct 2025 · Linux Foundation

Reducing MTTR and Alert Fatigue with PyTorch-Powered Anomaly Detection

PyTorch Conference 2025 — correlating multivariate telemetry to surface meaningful alerts with diagnostics.

May 2026 · apidays New York

Vibe Coding Has a Measurement Problem — Here's How to Fix It

Addressing observability and quality metrics for AI-generated code in production systems.

Sep 2026 · Forth Roadmap Conference

Closing the Last-Mile Gap: AI-Powered Electric Delivery for Healthcare Equity

How electric last-mile networks can extend to prescriptions, diagnostics, and medical supplies for underserved communities.

04

Service & mentoring

40+ manuscripts peer-reviewed across ~15 venues, plus startup-pitch and awards judging. Mentoring student teams at CSULB and UC Santa Cruz.

Journal reviewer

  • IEEE J. of Biomedical & Health Informatics
  • IEEE Access
  • JMIR — digital health

Conference PC / reviewer

  • AMIA — 8 papers
  • COMPSAC 2026 — 6 papers
  • HICSS — 4 papers
Let's build something that matters

Working on care, infrastructure, or AI at scale? Let's talk.

Open to research collaboration, co-authorship, advisory conversations, speaking invitations, and mentorship requests.