Forward Deployed Engineer · DevOps & AI Platform
Technical Manager at SugarCube Networks · Seoul, South Korea
I turn ambiguous AI and infrastructure needs into deployable tools, reliable platforms, and explicit operating paths.
- Work with users and operators to turn incomplete requests into explicit workflows and system boundaries.
- Build across applications, control planes, delivery, identity, and observability when the operating path requires it.
- Treat readiness, rollback, and recovery as part of product behavior—not as work left until after release.
- StreamGPU — FDE · AI Platform: Took a physical GPU rental idea through BMC and Docker validation to a Xen-based production service, connecting product discovery, implementation, release, and operations.
- Market Auditor — FDE · Operator Tooling: Turned a merchandiser's manual marketplace investigation into a workflow that separates read-only audits, operator approval, safe corrections, and post-write verification.
- Platform Foundations — DevOps: Built shared OIDC, self-hosted CI runners, registry-based delivery, canary and recovery gates, and centralized Prometheus, Loki, and Grafana operations.
- Traceknot — Developer Experience: An evidence-bound QA framework that connects coding-agent completion claims to executable evidence and deterministic verdicts.
- Previously worked as a Senior AI Engineer on an on-premises security-alert classification and serving system.
- Co-founded a Voice AI company and led its product application, data pipeline, GPU infrastructure, and operating environment.
Korean — Native · English — Conversational · Mandarin Chinese — Conversational