Fuchun Li
PDF Resume
Target role: Senior Consultant Specialist (AI Gateway)
15 yrs hands-on engineering · 7 yrs architecture leadership · Tencent Cloud Certified Architect / Hall of Fame
BornJan 1988| GenderMale| HometownChangning, Hunan| Expected pay30-45K RMB/mo
EducationBeihua University · Software Engineering · B.S. (Jul 2011)| EnglishCET-4 515 · professional reading/writing
Phone/WeChat13682320515| Email505847426@qq.com| LocationShenzhen (open to Guangzhou / Xi'an)
Fuchun Li
Role Fit
  1. AI / API gateway platform delivery: Built and operated LLM platforms with dual API gateway isolation (AgentX / DigitalHuman); multi-model LLM routing, tool calling, and RAG, covering design, development, test, deploy, and day-2 maintenance.
  2. SRE & incident response: Hands-on monitoring, alerting, triage, and recovery (Prometheus / ELK / CAT / Zabbix); root-cause fixes and preventive automation under 99.9%+ SLO targets.
  3. Gateway patterns: Multi-tenant isolation, authentication/authorization, rate-limit-style traffic control (Sentinel), and observability on open APIs; ready to deepen Kong / OpenResty plugin work on the job.
  4. Enterprise Kubernetes: Production Kubernetes / Docker / Istio / GitOps across PaaS and SaaS; ingress and service-networking experience via Nginx / mesh.
  5. Languages & async systems: Strong Java / Python for distributed services; productive with Golang for platform tooling; deep RabbitMQ and Kafka-class messaging for high-throughput async paths.
Skills Fit
  1. AI Gateway adjacency: Multi-provider LLM integration, prompt/tool pipelines, RAG, eval-minded iteration, and gateway-isolated service edges; ready to ramp on Portkey / APISIX / Higress.
  2. Platform programming: Java / Python concurrency and performance in distributed systems; Golang for utilities/automation; Nginx-fronted gateways in production (Zenjoy).
  3. Kubernetes in enterprise: Workload packaging, promotion, and day-2 ops for stateful/stateless services; cloud production ops (Tencent Cloud + private-cloud bank deliveries).
  4. Messaging & throughput: RabbitMQ / RocketMQ / NATS-style async systems; daily push volumes from millions to ~10M with reliability gates.
  5. Production support: User-facing troubleshooting, runbook discipline, and Tech Lead mentoring across 15 years of software engineering (5+ years platform/SaaS scale).
Experience
Jul 2025 – Jul 2026 | Microsoft Xiaoice (AI digital human / Agent) | Platform Lead / Enterprise Delivery
  1. Shipped DigitalHuman and AgentX on Docker/Kubernetes with dual API gateway isolation; cut end-to-end latency 2s → 500ms and held 99.9% availability for bank/insurance private clouds.
  2. Owned multi-model LLM routing via MCP / Tool Calling, RAG pipelines, and day-2 ops (logs/metrics/alerts) for platform reliability.
  3. Led Industrial Bank / Cathay / Ruizhong private-cloud deliveries with change-control discipline and production support under regulated workloads.
Nov 2024 – Jun 2025 | ZTOGO / Gofo Express (cross-border last-mile) | Backend Lead
  1. Led microservice delivery for last-mile logistics; scaled capacity from 100K → 1M daily parcels across five countries.
  2. Built high-throughput messaging (~10M pushes/day, 99.99% reach) and Flink + StarRocks visibility — SRE-style monitoring for delay and backlog risk.
  3. Established design review, code review, and on-call runbooks for cross-module incidents.
May 2023 – Nov 2024 | Keno (HeroDash global CX SaaS) | Architect
  1. Architected HeroDash on Spring Cloud + Kubernetes + GitOps; multi-tenant SaaS with open APIs, supporting 600+ concurrent sessions and 5M+ users.
  2. Integrated Nacos/Sentinel/Seata for policy-style traffic control and resilience; elastic capacity cut ops cost 50%+.
  3. Hardened bi-weekly release trains with observability for cluster health, queues, and API SLOs — production support for hybrid-cloud workloads.
Nov 2021 – Apr 2023 | Zenjoy (game publishing) | Java Architect
  1. Rebuilt high-traffic publishing platform (~1.4B users, peak ~30K QPS); containerized CI/CD on Tencent Cloud with minute-level rollouts.
  2. Deployed full-stack observability (ELK / Prometheus / Zabbix): seconds to detect, minutes to remediate production incidents.
  3. Shipped Nginx-fronted async WebFlux gateway with circuit-breaking clusters for blast-radius control.
Apr 2020 – Nov 2021 | Authine (Yunshu low-code PaaS) | Architect
  1. Built Yunshu’s cloud-native stack from scratch: Docker + Kubernetes + Istio + Spring Cloud for HA and horizontal scale.
  2. Stood up a 15-person CCB Beijing delivery team; three milestones to branch-scale rollout under bank change windows.
  3. Raised automated deployment coverage to 90% and shortened delivery cycles ~40%.
Sep 2018 – Apr 2020 | Lifesense (wearable health IoT) | Senior Engineer
  1. Owned multi-tenant health-cloud microservices and device ingest gateways; post-upgrade traffic hit 100M+/day at 99.99% availability.
  2. Ran Docker/Kubernetes + Jenkins CD with CAT/ELK tracing for Linux production support.
  3. Delivered Tianan Insurance health module; enterprise accounts grew to 300+.
Jan 2015 – Sep 2018 | Qianhai Image (cross-border supply chain) | Senior Java Engineer
  1. Built Dubbo microservices for order/warehouse/logistics/finance from scratch with RabbitMQ-backed async flows; sustained ~1M orders/day.
  2. Tuned end-to-end latency (order query 2.3s → 0.7s); introduced CI/CD and monitoring gates.
  3. Drove code review and knowledge sharing; improved team delivery throughput ~40%.
Jul 2011 – Jan 2015 | 7Road (browser games) | Java Engineer
  1. Owned core platform modules (account/billing/CS/ops) under Linux production load.
  2. Supported peak concurrency ~2,000 and ~200M lifetime users.
  3. Maintained CMS/SEO sites for live-ops campaigns and brand conversion.
Projects
AgentX Agent Engine | LLM routing / tool calling | Jul 2025 – Jul 2026

Context: Enterprise agent engine (Java + Python + Vue designer). Multi-source ingest, Agent orchestration, tool calling, and admin configuration — the control-plane work closest to an AI Gateway.
AgentX architecture

Highlights:

  1. Java Agent orchestration with multi-model LLM and MCP tool calling (sense → think → act).
  2. Python DAG Workflow + designer: multi-step tasks, node scheduling, state, RabbitMQ async execution.
  3. Docker/Kubernetes packaging with dual API gateway isolation; ship tool/rule updates without redeploying core business services.
Stack: Spring Boot + FastAPI + MyBatis-Plus + Sa-Token + Redis + RabbitMQ + ES + OSS + Celery + Docker + Kubernetes + LLM/MCP
DigitalHuman Platform | RAG / real-time LLM UX | Jul 2025 – Jul 2026

Context: Enterprise digital-human production and ops — LiveKit real-time, RAG Q&A, TTS/ASR, multi-engine streaming; bank/insurance private-cloud delivery.
DigitalHuman architecture

Highlights:

  1. End-to-end ASR/TTS/LLM/RAG/LiveKit path; 99.9% availability at ~10K-user scale.
  2. Multi-engine TTS + SSE streaming; end-to-end latency 2s → 500ms (75% reduction).
  3. Private-cloud standard edition with on-prem middleware/DB adaptation for Industrial Bank / Cathay / Ruizhong.
Stack: Spring Boot + MySQL + Redis + RabbitMQ + ES + Nacos + MinIO + Docker + Kubernetes + LLM/ASR/TTS/LiveKit
HeroDash Global CX SaaS | Open API / multi-tenant | May 2023 – Nov 2024

Context: Hybrid-cloud contact-center SaaS (work.herodash.ai) — tickets, knowledge, AI assist, open APIs.
HeroDash architecture

Highlights:

  1. Spring Cloud on Kubernetes with GitOps; 600+ concurrency, 5M+ users.
  2. Sentinel/Nacos/Seata for traffic governance and resilience; ops cost down 50%+.
  3. Production monitoring for cluster health, messaging, and API SLOs in a multi-tenant model.
Stack: Spring Cloud + Nacos + Sentinel + Seata + MySQL + PostgreSQL + Redis + ES + RabbitMQ + NATS + Kubernetes
Yunshu Low-Code PaaS | Cloud-native platform | Apr 2020 – Nov 2021

Context: Domain-driven low-code product. Metadata-driven app builders for enterprise and bank rollouts.
Yunshu architecture

Highlights:

  1. From-scratch path: Docker + Kubernetes + Istio + Spring Cloud for high availability and scale-out.
  2. SDK packaging, non-functional baselines, and bank delivery playbooks.
  3. Built a 15-person CCB Beijing team; three milestones to branch-scale production.
Stack: DDD + COLA + Spring Boot + Spring Cloud + Kubernetes + Docker + Istio + Vue
Zenjoy Publishing Platform | Nginx gateway + observability | Dec 2021 – Apr 2023

Context: Core cash flow / publishing stack — login, payments, ads, and unified ops portals at game scale.
Zenjoy architecture

Highlights:

  1. ~1.4B users / ~9M DAU / peak ~30K QPS; Tencent Cloud containers + GitLab CI/CD.
  2. Nginx + WebFlux async gateway with circuit-breaking clusters.
  3. ELK / Prometheus / Zabbix: seconds to detect, minutes to clear incidents.
Stack: Dubbox + WebFlux + MySQL + Redis + ZooKeeper + Apollo + Nginx + Tencent Cloud containers
Targets
What I look for
  1. Sustainable business with real production load
  2. Healthy engineering culture and clear ownership
  3. Modern AI platform / gateway / cloud-native stack
Tech advocacy / media
Tencent CloudTencent Cloud
ToutiaoToutiao
WeChatWeChat
CSDNCSDN
GitHubGitHub