Projects
These projects reflect the systems I have been building: release checks, support automation, incident triage, model-quality gates, drift monitoring, SLO alerting, and Kubernetes operations. The repositories include local tests, CLI or API behavior, and deployment-readable infrastructure assets.
iOS | SwiftUI | SRE | Swift Testing
I built a native iOS and iPadOS SRE incident commander app with deterministic service metrics, tested SLO/error-budget calculations, runbook-driven incident workflow, Reliability Lab simulations, and Markdown post-incident review export.
MLOps | Feature Quality | FastAPI | Kubernetes
I built a feature-publication gate that checks contracts, drift, freshness SLAs, and accountable ownership before feature data reaches model scoring or training workloads.
RepositoryAI Automation | Incident Response | Redaction | Evaluation
I built an AI-assisted runbook generator with shared CLI and API logic, redaction-aware review, and quality-history gates that make response drafts easier to evaluate before use.
RepositoryLLM Evaluation | Release Gates | FastAPI | Prometheus
I built a deterministic evaluation gateway that compares baseline and candidate results, tracks regression budgets across history, and exposes CLI, API, and Markdown review paths for release decisions.
RepositoryRelease Reliability | Python | FastAPI | Prometheus
I built a deterministic release-gate service that evaluates rollout evidence, customer impact, latency, saturation, and incident signals to recommend continue, pause, or rollback decisions.
RepositoryAI Support | Vector Search | FastAPI | Evaluation
I built a deterministic support assistant that retrieves the best incident runbook, returns cited next actions, and includes retrieval evaluation gates for support-quality checks.
RepositorySRE | Alert Correlation | FastAPI | Prometheus
I built a deterministic SRE engine that groups noisy alerts, suppresses duplicate pages, and checks ownership, escalation, and runbook readiness before incidents reach responders.
RepositoryCI/CD | Supply Chain | FastAPI | Prometheus
I built a release gate that evaluates deployment readiness and software-supply-chain evidence, including SBOMs, provenance, signatures, vulnerabilities, and licenses, through shared CLI and API workflows.
RepositoryMLOps | Canary Release | FastAPI | Kubernetes
I built a model-serving platform with deterministic canary routing, shadow evaluation, Prometheus metrics, and deployment assets that demonstrate safer rollout review.
RepositoryTerraform | Policy as Code | FastAPI | CI/CD
I built a policy engine that reviews Terraform plan JSON for cost, destructive changes, public ingress, tag hygiene, and privileged IAM before infrastructure changes proceed.
RepositoryMLOps | Model Monitoring | Prometheus | Terraform
I built a model-monitoring gate that combines input-drift checks with delayed ground-truth accuracy and macro-F1 history before promotion.
RepositorySRE | Prometheus | SLO Burn Rate | API
I built an SRE lab that evaluates multi-window burn rates, routes page/ticket/ok decisions, and writes incident-ready reports for platform reliability practice.
RepositoryKubernetes | Log Triage | Runbooks | Metrics
I built a Kubernetes-focused triage service that scores log batches, identifies likely incident classes, returns runbook steps, and exposes Prometheus metrics.
Repository