Automating Reliability | Securing Cloud Infrastructure | scaling AI Operations
I engineer production-grade cloud platforms that are secure by default and self-healing by design. Currently at Pipeline Inc. (Tokyo), I focus on the intersection of Infrastructure as Code, Identity-Based Security, and Observability.
Real-world systems I am currently building and operating.
Designed and deployed an RBAC-based SSH access system using Teleport, replacing static keys for 24+ production nodes.
- Tech: Teleport, SSO, IAM Roles, Linux Hardening.
- Outcome: Eliminated static credential management and enforced audit trails for all server access.
Orchestrating data pipelines and supporting Vector DB + RAG integration for internal AI tooling.
- Tech: Apache Airflow, Vector Databases, Python, AWS S3.
- Outcome: Optimized infrastructure for high-throughput LLM queries and automated data workflows.
Built a centralized monitoring stack to reduce "Mean Time to Detect" (MTTD).
- Tech: Prometheus, Grafana, CloudWatch, Loki.
- Outcome: Reduced incident detection latency by ~85% through proactive alerting rules.
| Domain | Technology Ecosystem |
|---|---|
| Cloud Infrastructure |    |
| IaC & Config Mgmt |    |
| Orchestration |    |
| CI/CD & GitOps |    |
| Security (DevSecOps) |    |
- Immutable over Mutable: Servers should be cattle, not pets. If it breaks, replace it; don't patch it.
- Identity is the New Perimeter: IP whitelisting isn't enough. I enforce access via IAM Roles and Short-lived Certificates.
- Observability is Testing in Production: If you can't see it, you can't fix it. Logs tell you what happened; metrics tell you why.