«yourhandle»

«Your Name»DevOps Engineer · Dasmariñas, Philippines

Built to survive the spike.

I keep online gaming and fintech platforms up when traffic triples in ninety seconds. Infrastructure, live streaming, test automation, and applied AI — run as one system, not four job titles.

Service healthAll operational
svc/infrastructureInfrastructure42ms p99
svc/streamingLive streaming1.8s glass-to-glass
job/qualityQuality gates96% suite pass
svc/intelligenceApplied AI320ms ttft
Trailing 30d · no unplanned downtime

svc/ — four services, one operator

What I run

These aren't separate jobs. The streaming stack needs the pipeline, the pipeline needs the test suite, and the AI layer reads what all three produce.

svc/infrastructure

Infrastructure & platform

Design and run the AWS footprint behind high-traffic gaming and payment workloads, with deploys that don't need a maintenance window.

  • AWS EC2 / RDS / ElastiCache
  • Terraform + Terragrunt
  • Kubernetes + Helm
  • ArgoCD
  • GitHub Actions
  • Cloudflare WAF
  • Prometheus + Grafana
  • Loki
svc/streaming

Streaming systems

Build low-latency delivery paths from ingest to player, plus the auth layer that stops stream keys leaking.

  • MediaMTX
  • Wowza
  • WebRTC / coturn
  • LL-HLS
  • RTMP / SRT
  • JWT + Redis revocation
  • S3 recording
  • ffprobe health checks
job/quality

Quality engineering

Turn release confidence into something measurable — end-to-end coverage, load profiles modelled on real tournament spikes.

  • Playwright
  • k6 load & spike tests
  • Postman / curl
  • API contract tests
  • Auth & rate-limit suites
  • CI regression gates
svc/intelligence

Applied AI

Put LLMs where they pay for themselves: incident summaries, anomaly triage, and admin tooling that reads logs so nobody has to.

  • Anthropic Claude API
  • Structured JSON output
  • Tool use / function calling
  • RAG + pgvector
  • Prompt evaluation
  • Log & anomaly summarisation

releases/ — newest first

Selected deployments

Every entry shipped to production and carried real traffic. Numbers are measured, not estimated.

v3.1.02026-04

Tournament-scale live streaming

Rebuilt the ingest-to-playback path on MediaMTX with LL-HLS and signed playback tokens. Held «40,000» concurrent viewers through a finals weekend with no manual intervention.

  • MediaMTX
  • LL-HLS
  • Redis
  • Cloudflare
«−62%» startup latency
v2.4.02025-11

Zero-downtime release pipeline

Moved «12» PHP services from hand-run deploys to blue/green behind an ALB, with automated schema migration and a one-command rollback.

  • GitHub Actions
  • Terraform
  • ALB
  • MySQL
«18 min → 4 min» deploys
v1.8.02025-06

Payment anomaly watch

A Claude-backed service that reads transaction and auth logs, clusters suspicious patterns, and files a structured summary into the on-call channel.

  • Claude API
  • Structured output
  • PHP
  • Redis
«first-hour» fraud detection
v1.2.02025-01

Regression gate

Playwright suite over the full auth and cashier flow, plus k6 spike profiles modelled on real tournament traffic. Nothing merges without both going green.

  • Playwright
  • k6
  • CI
«96%» critical-path coverage

history/«6» years in production

How I got here

«2024 — present»

«Senior DevOps Engineer» · «Company»

Own the production platform for «live casino and sportsbook» traffic: infrastructure, streaming delivery, release pipeline, and on-call.

«2022 — 2024»

«DevOps Engineer» · «Company»

Migrated the LAMP estate onto managed AWS services and introduced infrastructure as code across «three» environments.

«2020 — 2022»

«Systems / Support Engineer» · «Company»

Ran Apache, PHP-FPM, and MySQL in production. Learned what breaks at 2am, which is most of what I do now.

Open to new work

Let's talk about your uptime

Platform builds, streaming delivery, or a release process that keeps waking people up. Tell me what's breaking.

«you@example.com»