Software Engineer & SRE: I design, build and run production systems. 12+ years keeping critical platforms up.
🟢 Open to work: Software Engineer (backend / platform) or SRE, remote or Santiago, available immediately.
I write the services and I run them in production: Go, Rust and TypeScript backends with hexagonal architecture and resilience patterns, plus cloud infrastructure, Kubernetes, IaC, CI/CD and observability. I've worked on AWS since 2019 and have led single-region to multi-region migrations in production. I hold 11 CNCF certifications.
| Project | Stack | What it does |
|---|---|---|
| orderflow | Go · NATS JetStream · PostgreSQL · Stripe · Auth0 · React/TS · OpenTelemetry · Helm · Terraform | An e-commerce order, payment and fulfillment platform built like production: five hexagonal services, a saga with compensations, a transactional outbox and an idempotent inbox, and a circuit breaker, retries with jitter, bulkheads, rate limits and idempotency keys. It includes a Stripe adapter with verified webhooks, an Auth0/JWT backoffice with RBAC and an audit log, Grafana SLOs and burn-rate alerts, and chaos scripts that show the saga compensating. |
| orderflow-platform | Terraform · AWS (EKS, Karpenter, Aurora Global, Route 53, WAF) · Argo CD · Argo Rollouts · KEDA · Kyverno · Chaos Mesh · k6 · Prometheus | How orderflow runs in production: multi-region active/passive AWS in Terraform, GitOps with Argo CD, canaries that abort themselves on SLO breach, autoscaling on queue depth, policy as code and a signed supply chain. make prove runs a chaos game day on kind: 99.9% availability while pods, NATS and Postgres were killed (DB writable again in 1.2 s), and a bad release rolled back with no human. |
| hookrelay | Rust · axum · tokio · SQLite | A reliable webhook relay. It verifies HMAC X-Hub-Signature-256 signatures (WhatsApp/Meta style), persists events to a durable queue, and delivers them with retries, exponential backoff, a dead-letter queue, idempotency keys and Prometheus metrics. |
| slo-probe | Go · Prometheus | A synthetic prober that computes SLIs, error budgets and multi-window burn-rate alerts. It exports metrics, and its CLI report can serve as a deploy gate. |
| hookrelay-console | TypeScript · React · Vite | An operator UI for hookrelay with deliveries, the dead-letter queue, one-click replay and a mock mode. |
| pasito | TypeScript · React · Vite · PWA · IndexedDB | An offline-first learning app for kids aged 2–7 that teaches speaking, numbers and reading (Spanish phonics). It uses the Web Speech API for text-to-speech and speech recognition, spaced repetition and adaptive difficulty, and has a parent dashboard with streaks. The domain is hexagonal and tested, with Playwright and axe e2e tests. No ads, no tracking. Try it live |
| kernel-patch-validator | Shell | Checks Linux kernel patches before you submit them upstream. I announced it on linux-kernel-mentees, where Greg Kroah-Hartman reviewed it ("I like the idea") and suggested moving the checks into checkpatch.pl, which I then proposed as checkpatch patches. |
| Copihue (formerly Chronos) | C · x86-64 Assembly | An experimental systems language (not production-ready yet). Its compiler, written in C, emits x86-64 assembly and native ELF64 binaries with no libc, plus a freestanding mode used by the AtomicOS kernel. |
| AtomicOS | Assembly | An operating system written from scratch in Assembly. |
| leafwise | Go · PyTorch · ONNX Runtime · React PWA · Postgres · MinIO | A smart gardening assistant. It schedules watering and feeding from a soil water balance driven by real Open-Meteo weather, and diagnoses leaf diseases from photos with a MobileNetV3 I fine-tuned on PlantVillage (94% test accuracy), served through ONNX Runtime behind a hexagonal diagnosis port with an optional vision-LLM second opinion. |
- Linux kernel (2025): got cleanup patches accepted into Linux staging by Greg Kroah-Hartman.
- Gauss Control (2026): owned a multi-tenant IoT platform on GKE end to end, with Kafka, Redis, BigQuery, Terraform Cloud, Python and Go, at USD 50K+ a month in cloud spend.
- Macal, Head of Operations (2021–2025): built the SRE team from scratch and took AWS from single-region to multi-region. Cut infrastructure costs by 40% and reached 99.9% availability.
- COPEC (2019–2021): migrated 200+ applications to AWS, shipped 13 deploys a day with zero downtime, and ran them multi-region on Aurora Global Database, RDS cross-region replicas, DynamoDB Global Tables and CDC, with Route 53 failover and regular disaster-recovery drills.
- Itaú, Practia (2011–2019): built QA automation for banking and enterprise systems.
- Languages: Go, Rust, TypeScript, Python, Shell
- Cloud: AWS (multi-region, Route 53, Global Accelerator, EKS/ECS, Aurora, RDS, DynamoDB), GCP (GKE, BigQuery), Cloudflare
- Platform: Kubernetes, Docker, Kafka, Redis, Argo CD
- IaC and CI/CD: Terraform, CloudFormation, GitHub Actions, GitLab CI, Jenkins
- Data: PostgreSQL, MySQL, SQL Server, DB2, MongoDB, Redis
- Observability: Prometheus, Grafana, OpenTelemetry, New Relic, CloudWatch
- Certifications: 11 CNCF certifications, MBA
LinkedIn: https://www.linkedin.com/in/ignacio-pe%C3%B1a-2905b454/
Native Spanish speaker, professional working proficiency in English.
11 Linux Foundation / CNCF certifications: CKAD · KCNA · KCSA · PCA (Prometheus) · OTCA (OpenTelemetry) · CAPA (Argo) · CGOA (GitOps) · CNPA (Platform Engineering) · CBA (Backstage) · KCA (Kyverno) · CCA (Cilium)