diff --git a/RESUME.txt b/RESUME.txt new file mode 100644 index 0000000..8be57de --- /dev/null +++ b/RESUME.txt @@ -0,0 +1,228 @@ +================================================================================ + ROCK LIANG - RESUME +================================================================================ + +CONTACT INFORMATION +Email: rock@riotpiao.com +Portfolio: https://riotpiao.com +GitHub: https://github.com/rockliang + +================================================================================ +PROFESSIONAL SUMMARY +================================================================================ + +Experience: + 6+ years building production-grade infrastructure, backend systems, and LLM + applications. Specialized in Kubernetes orchestration (Talos Linux, ArgoCD), + Infrastructure-as-Code (Terraform), distributed systems design, AWS scale + operations, and GPU-accelerated inference. + +Core Competencies: + Production Kubernetes (Talos, GitOps/ArgoCD, 99.2% uptime). Terraform IaC + with drift detection. AWS distributed systems (57+ regions). Golang backend + services. Observability (Prometheus/Grafana/Loki/Tempo). Message queuing + (Kafka/Temporal workflows). GPU inference (vLLM, 60% latency improvement). + Zero-trust networking (OIDC/RBAC). + +Strengths: + Owning systems end-to-end—design doc through oncall. Building fault-tolerant + platforms with 99.2% automation. Unblocking teams with clear technical + communication. + +================================================================================ +EXPERIENCE +================================================================================ + +--- +SELF-HOSTED PRODUCTION (riotpiao.com) +DevOps / SRE / SDE +Home Lab | May 2025 — Present + +Challenge & Solution: + Built AWS from scratch at home on bare-metal Kubernetes—Talos Linux for + secure immutable OS, ArgoCD for GitOps pipeline automation, SOPS for + encrypted secrets management, cert-manager for PKI. Result: + production-grade multi-tenant platform with Kafka, PostgreSQL, S3, GPU LLM + inference, Temporal workflows in 4-node cluster. + +Tech: + • Kubernetes (Talos, ArgoCD) — Talos immutable, atomic updates, no SSH, full + GitOps state. ArgoCD automatic sync with drift detection. Cilium eBPF CNI. + • Terraform + Kustomize — Declarative infrastructure. State management via + Terraform Cloud. Kustomize overlays for multi-env. + • Secrets & Security — SOPS age encryption with git audit trail. Authentik + OIDC SSO with Kubernetes RBAC. cert-manager + Let's Encrypt (30d renewal, + 7d pre-expiry alerts). + +Infrastructure & Services: + • Longhorn 3-replica block storage. MinIO S3-compatible API. CloudNativePG + PostgreSQL with pgvector for embeddings. vLLM GPU inference (Qwen3-32B, + 60% latency vs baseline). + • Observability stack: Prometheus metrics, Grafana dashboards, Loki logs, + Tempo traces, OpenTelemetry. 99.2% uptime via monitoring automation. + • Kafka/Redpanda streaming. Temporal workflows for fault-tolerant jobs. + Forgejo CI/CD with Docker registry. DinD runners for parallel builds. + +--- +ROYAL BANK OF CANADA (RBC) +Lead Software Engineer +Toronto, ON | Nov 2024 — May 2026 + +Challenge & Solution: + Deployments were manual and slow—teams blocked 2+ hours per release waiting + for provisioning. Standardized on Terraform Cloud for idempotent IaC. Built + Temporal workflows for multi-cloud orchestration with built-in retry logic + and idempotency keys. Reduced deploy time from 2 hours to 20 minutes, + achieved 99.2% automation. + +Drift Detection & Reconciliation: + • Built Kubernetes CronJob (Golang) to continuously detect and reconcile + Terraform state drift against actual infrastructure. Automatically + detects divergence (network policies, security groups, storage) and + triggers remediation without manual intervention. + • Coordinated across 4 platform teams, unified 200+ microservices into + single IaC standard. Cut onboarding time for 12 teams. + +Notifications & SLA: + • Designed Slack notification service with Golang workers. Alerts on apply + failures, drift detection, provisioning bottlenecks. Operators resolve + issues in <5 minutes via notification-driven workflow. + +Stakeholder Alignment: + • Led requirement gathering across 4 platform teams. Translated technical + decisions for non-technical stakeholders. Secured buy-in for platform + migration. Unblocked 3 stalled projects by disambiguating specs. + +Tech: + • Golang, Kubernetes, Terraform Cloud, IaC, Distributed Systems + • Slack API, multi-cloud orchestration, state drift reconciliation + • Notification-driven operators, 99.2% automation + +--- +AMAZON WEB SERVICES (AWS) +Senior Software Engineer (Step Functions) +Seattle, WA | 2022 — 2024 + +Challenge & Solution: + Owned Distributed-Map from design doc to production launch across 57+ + regions. Under tight deadline, introduced separate state in Map vs new + functions—simplified customer adoption. Built checkpoint recovery for + mid-workflow failures so customers resume without full re-run. + +Distributed Systems Engineering: + • Solved critical edge cases: race conditions in concurrent state updates, + dependent service failures, message deduplication, exactly-once semantics + across distributed workers. + • Introduced JSON state input for larger payloads—unlocked new customer use + cases (batch processing, ML model serving, data pipelines). + +Production Ownership & Reliability: + • Full oncall ownership—built CloudWatch dashboards, wrote operational + runbooks, debugged production issues live. Achieved sub-100ms P99 latency + across 57+ regions and 20x burst traffic scaling. + • Disaster recovery procedures and incident response. State management for + failed/retried jobs. + +Tech: + • Java, AWS Step Functions, DynamoDB state storage, gRPC, Distributed Systems + • CloudWatch observability, ownership, disaster recovery, exactly-once delivery + • 57+ regions, sub-100ms P99, 20x burst traffic + +--- +NETFLIX TITUS (Container Orchestration) +Software Engineer Intern +Los Gatos, CA | May — Aug 2019 + +Challenge & Solution: + Streamlined Personal Data Detection (PDD) to identify classified data + exfiltration anomalies—achieved 97.8% accuracy with minimal false positives. + Built fault-tolerant Golang connector for PDD integration with 28% p99 + latency improvement over legacy implementation. + +Compilation & Integration: + • Re-integrated SmartRegex with CMake & C++ on Linux/Unix platforms. + Achieved 5x faster deployment through automated build pipeline and + dependency resolution. + • Prevented classified data leakage across Netflix's distributed workload + platform. + +Tech: + • C++, CMake, Golang, Docker, pattern matching, fault tolerance + • Build automation, 97.8% detection accuracy, 28% latency improvement + +--- +NAV CANADA (Air Traffic Control Systems) +Summer Student +Ottawa, ON | May — Aug 2018 + +Challenge & Solution: + Maintained enterprise CFPS web application in Agile development. Built + Django NOTAMJ polls app to meet iteration goals. Improved deploy stability + with Sonar code coverage analysis—reduced defect escape rate. + +Frontend & User Experience: + • Created FWGS weather briefing interface with ReactJS for Air Traffic + Control. Real-time meteorological data visualization for pilot briefings. + • Deployed to Canadian airspace (50+ airports). Improved ATC + decision-making with cleaner weather briefing UX. + +Tech: + • ReactJS, Django, Agile methodology, Sonar code coverage + • Weather data visualization, pilot decision support systems + +================================================================================ +TECHNICAL SKILLS +================================================================================ + +Languages: + Golang, Java, Python, C++, JavaScript/TypeScript, SQL, Bash + +Cloud & Infrastructure: + Kubernetes (Talos Linux, k3s, ArgoCD GitOps) + Terraform (IaC, state management, drift detection) + Docker, OpenShift, AWS (DynamoDB, Step Functions, CloudWatch) + +Distributed Systems: + Temporal workflows, Kafka/Redpanda streaming + gRPC, message queuing, exactly-once semantics + State management, disaster recovery, idempotency + +LLM & AI: + vLLM (GPU inference, Qwen3-32B, 60% latency improvement) + KServe orchestration, Ollama, TEI embeddings + Vector databases (pgvector), model serving + +Observability & Monitoring: + Prometheus (metrics), Grafana (dashboards) + Loki (log aggregation), Tempo (distributed tracing) + OpenTelemetry, CloudWatch, alerting & runbooks + +Networking & Security: + Authentik OIDC SSO, Kubernetes RBAC + cert-manager TLS, SOPS encryption + Cloudflare Tunnel (zero-trust), network policies + PKI, service-to-service auth + +Data & Storage: + PostgreSQL (CloudNativePG, pgvector) + DynamoDB, MinIO S3, Longhorn block storage + Data modeling, exactly-once consumption + +CI/CD & DevOps: + Forgejo Actions, GitHub Actions + Docker container registry, Kustomize + GitOps, declarative infrastructure + +================================================================================ +EDUCATION +================================================================================ + +University of Ottawa, Ottawa, Ontario +Master of Science in Computer Science (Software Systems) +2020 — 2021 + +University of Ottawa, Ottawa, Ontario +Bachelor of Science in Computer Science (Honours) +2016 — 2020 + +================================================================================