Architecting Autonomy

Harness Engineering

Kylin Miao

AI Agent Engineer @ Tencent WeCom

Evolved through Prompt → Context → Harness Engineering.
Now orchestrating multi-agent systems that work while I sleep.
Passionate about LLMs & Agents. The future is here, and I'm fully embracing it.

LLM / RAG

Core Stack

Multi-Agent

Architecture

Full-Stack

Development

[ Execute ]

~/projects/selected_processes

Monitoring core background tasks from Tencent anti-fraud cluster and independent edge deployments.

DEPLOYED

PID: 00001

~/betaline

全栈独立开发上线，集成 AI 视频分析与岩点检测

EDGE_ACCEL■■■■■■■■■■■■300x

DETECTION■■■■■■■■■■■■92%

▸ [SYS] Compressing model for CoreML...▸ [OK] Edge inference latency dropped to 12ms.

~/openmemory-plus

AI Agent 双层记忆框架，实现跨会话 95%+ 上下文保持率

RETENTION■■■■■■■■■■■■95%+

REDUNDANCY■■■■■■■■■■■■-40%

▸ [INFO] Compacting redundant memory vectors...▸ [OK] Context retention stable at 95%+.

./anti-fraud-governance

多维度算法驱动的反作弊治理系统，年化收益超300万。

BANNED_SVCS■■■■■■■■■■■■820+

ANNUAL_REV■■■■■■■■■■■■3M+

▸ [SYS] Executing Label Propagation...▸ [OK] 820+ malicious providers isolated.

社区检测算法关联图谱

[ Execute ] ->

> list --all-processes

System Ready

Let's Build Something Together

Explore my complete portfolio or learn more about my background and expertise.

[ View_Projects ][ About_Me ]