Owen Villareal

Nutrition tracking app built mobile-first with natural language food logging, LLM-powered insight engine, and deep macro/micronutrient trend analytics. Features barcode scanning, NLP text parsing via LLM extraction, self-hosted USDA + Open Food Facts food database with pg_trgm fuzzy search, weight tracking with goal projection, and a nightly AI insights pipeline surfacing personalized dietary patterns.

URL shortener built in Go with Redis caching, rate limiting, and structured observability — containerized with Docker builds, deployed to Kubernetes with HPA, and monitored end-to-end with Prometheus, Grafana, and distributed tracing. Include full-stack web app capabilites with auth flow and user specific tooling for signed vs anon usage.

Event-driven distributed monolith simulating an ecosystem of animals, plants, and weather patterns as the domain to generate big data and explore architectural patterns. Built with a polyglot stack (Go, Elixir, Python) communicating via Kafka, with Redis and PostgreSQL for state management, Docker for containerization, and Grafana for observability. A sandbox for testing out distributed systems concepts like event sourcing, CQRS, stream processing, and microservices vs monolith tradeoffs. Endgoal is a data platform architecture pattern showcase and a YouTube series documenting the design and development process.

Distributed task queue built from scratch in Go with Kafka-backed priority lanes, exponential retry, dead letter queue, and a custom Kubernetes autoscaler that scales worker pods based on consumer lag rather than CPU. Obervability and logging via Grafana, Promethues, Loki. Production grade with CI/CD pipelines.