CV (AI Application Development Focus)

Personal Information

Period	Degree	School	Major
2006.9 – 2010.6	Bachelor	Changchun Institute of Technology	Electrical Engineering & Automation

14 years of experience in infrastructure and platform engineering across manufacturing, finance, telecom, and internet sectors; recently focusing on AI + Operations/Business applications.
Hands-on expertise in end-to-end AI application engineering: data ingestion → retrieval/indexing → RAG & Agent design → API/service deployment → observability & canary release → GitOps/CI/CD.
Strong in Kubernetes, GitHub Actions, GitOps, Terraform, Pulumi, bridging LLM applications with cloud-native platform engineering.
Practical experience in AIOps and observability: leveraging eBPF / OpenTelemetry / pgvector + LLM for root cause analysis, Runbook Q&A, alert deduplication, and automated incident response.

LLM & Agent
- Frameworks: LangChain / LlamaIndex / OpenAI Assistants API / function calling
- Agent Design: tool orchestration, retrieval chains, memory management, task planning & execution
- Models & Services: OpenAI / Claude / Qwen, Llama, Ollama, vLLM, Text & Embeddings APIs
RAG & Retrieval
- Embeddings & Rerank: bge-m3 / e5 / common rerankers
- Vector Databases: PostgreSQL + pgvector, Elasticsearch (kNN + BM25), Qdrant, Milvus
- Document Engineering: segmentation, metadata indexing, hybrid retrieval strategies
MLOps / AIOps
- Data pipelines: ETL, nearline ingestion of logs/metrics/traces into knowledge bases
- Evaluation: retrieval recall, answer accuracy, latency & cost monitoring, regression testing for prompts
- Governance: prompt/versioning, safety guardrails, rollout/rollback strategies
Platform & Automation
- Cloud-Native: Kubernetes, Helm, ArgoCD, FluxCD
- CI/CD: GitHub Actions, GitLab CI, Jenkins (AI app build, evaluation, release pipelines)
- IaC: Terraform, Pulumi (AWS/GCP multi-cloud)
Programming & Backend
- Python, Shell (familiar with Go, Rust, JavaScript)
- FastAPI / Flask / gRPC / WebSocket, task queues (Celery/Redis)
- Observability: Prometheus, Grafana, OpenTelemetry, DeepFlow

Scenario: Unified Runbook/alerts/change logs/monitoring docs into a retrieval-based Q&A system for SREs.
Tech: LangChain, pgvector, bge-m3, BM25 hybrid retrieval, ArgoCD read-only API, Grafana API.
Pipeline: GitHub Actions for daily incremental indexing; PR-triggered regression tests.
Impact: Reduced mean time to resolution by 30%+; average query response < 1.5s (cache hit).

Scenario: Query Terraform modules, Helm values, FAQs and auto-generate MR drafts.
Tech: LlamaIndex + pgvector, schema-constrained chunks, GitHub Actions PR Bot.
Impact: Reduced MR preparation time from hours to minutes.

Scenario: Natural language to observability query (metrics, traces, logs).
Tech: Toolformer-style router, functions like query_metrics(), search_traces(), get_logs().
Impact: Enabled first-line engineers to self-serve context queries, reducing SRE load.

ASKAI (Live Demo): https://www.svc.plus — conversational AI assistant with tool orchestration.
XScopeHub (AIOps Suite): Vector/OpenTelemetry/Postgres integration for nearline ETL & retrieval, with evaluation scripts.
XCloudFlow / XConfig: CI/CD + IaC modular framework, supporting canary/rollback for LLM apps.
Navi (Desktop AI Assistant): Cross-platform agent with guided tasks, knowledge Q&A, and memory-enabled tool orchestration.

Stack: Linux, Kubernetes, DeepFlow, NPM & APM, eBPF, Hybrid Cloud
Responsibilities: On-site deployment, tuning, and ops of observability systems; AI Agent PoC delivery.
Key Achievements:
- Optimized DeepFlow collector performance for heavy traffic (SynTao Finance).
- Delivered AI Agent PoC for COSCO Shipping (containerized deployment & QA testing).
- Supported observability deployments for energy & telecom clients (Shanghai Power, Zhejiang Mobile, Jiangsu Telecom).

Stack: Linux, Kubernetes, GitHub Actions, GitOps (ArgoCD), Ansible, Helm, Jenkins
Achievements:
- Developed Jenkins pipelines for industrial control (Ignition) automated deployment/upgrade.
- Built GitHub Actions + ArgoCD pipelines for ITSM & DMP systems, enabling full automation.
- Automated Grafana dashboards & alerts with Ansible Playbooks (Infra as Code).

Stack: AWS/AliCloud, Terraform, Pulumi, GitLab CI, GitOps (FluxCD)
Highlights:
- Re-architected AWS EKS Terraform modules (migration, upgrade, addon mgmt).
- Delivered Syngenta DevOps platform (EKS + GitLab) with IaC & app pipeline templates.
- Built multi-cloud monitoring & IAM integration with Keycloak.

Focus: Hybrid cloud design, large-scale Kubernetes migration, CI/CD modernization.
Projects:
- GrowingIO: Designed LB migration with automated load testing via Ansible.
- Blockchain client: Automated creation of 1000+ VMs via Python + API for elastic compute.

Focus: PaaS upgrades, DevOps pipelines, containerized platform delivery for finance clients.
Achievement: Led 6 major upgrades of China Everbright Bank’s CPaaS container platform.

Focus: Deepin Server OS V15 development, automated packaging & CI/CD pipeline.
Achievement: Reduced release cycles by 30% via automation scripts.

Focus: IT infra ops, CDN optimization, automated monitoring with Saltstack/Nagios.