fullstackjam — selected workRésumé
 Jam Ma — résumé fullstackjam.com

SRE · Platform · Full-stack

I keep systems alive, and ship what people actually touch.

SRE / Platform engineer with full-stack range. Five years across fintech and consumer hardware — Citi, then Huami.

↓70%
Model cost
100+
Services on GitOps
99.95%
Core service SLA
↓80%
Alert MTTR
Download PDF · 中文 GitHub Email Open to new opportunities

Experience

Huami · Zepp Health

SRE · DevOps

Nanjing · 2025.09 → now

  • Replaced GitLab's script-based delivery with Argo CD and a real GitOps model — git is the source of truth, 100+ services reconciled. Swapped random-10% canaries for header-based routing, so releases verify safely even without a staging environment.
  • Pushed the BU toward an AI-native workflow: rolled out Claude Code and Codex, and built a self-hosted Codex reverse proxy that cut monthly model spend by ~70%.
  • Rebuilt observability from availability-only to business-level metrics, with automatic alert diagnosis — MTTR down ~80%.

Citi

SRE · Tech Lead

Shanghai · 2022.08 → 2025.07

  • Led an in-house AIOps platform (Python / FastAPI) on OpenShift — a DAG workflow engine, NLP + GenAI command handling, and token-bucket / PID concurrency control. Throughput +300%, task failures −85%, 90% of common ops automated.
  • Owned delivery and reliability end-to-end: Tekton / Helm pipelines with staged rollout, Ansible config-drift detection, K8s + ELK + Grafana monitoring. 99.2% deploy success, −55% prod incidents, core SLA held 99.95%+.

Qunzhimai

SRE

Shanghai · 2021.03 → 2022.06

  • Built the team's first in-house ops platform from scratch (Flask + Aliyun SDK + Docker) — folded repetitive ops into one place and made the reporting data trustworthy.

Skills

Cloud & containers

Aliyun / Tencent Cloud / AWS · Kubernetes (HA, troubleshooting) · Docker · OpenShift

CI/CD & GitOps

Argo CD · Jenkins / Tekton · GitHub Actions / GitLab CI · Ansible · Terraform · Helm

Observability & reliability

Prometheus + Grafana + ELK · business metrics · SLO · automated alert diagnosis

AI engineering

Claude Code / Codex rollout · self-hosted Codex proxy · workflow redesign

Systems & dev

Linux tuning & troubleshooting · Nginx / Redis / MongoDB / OracleDB · Python (FastAPI / Flask) · Shell

Selected work

I build in the open — a cluster I can rebuild from zero, a one-command Mac setup, an AI agent living inside Lark.

Full case studies on the home page

Education

Jiangsu University of Science and Technology

B.E., Computer Science

2016 → 2021