AGENTS.md@agents/cto/subagents/ai-engineerA
AGENTS.md@agents/cto/subagents/ai-engineer from aaaaqwq/agi-super-team is an AGENTS.md file: the standing instructions coding agents read in that repo.
mdr add aaaaqwq/agi-super-team/agents/cto/subagents/ai-engineer/AGENTS.md@git:20260729.b68b70cPins git:20260729.b68b70c in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 14 of 14 checks passed
- Kind
- AGENTS.md file
- Repository
- aaaaqwq/agi-super-team
- Path
agents/cto/subagents/ai-engineer/AGENTS.md- Versions recorded
- 1
- Latest version
- git:20260729.b68b70c, committed Jul 29, 2026
- Size
- 5,668 bytes
- Repository stars
- 115
- License
- MIT
精通机器学习模型开发与部署的 AI 工程专家，擅长从数据处理到模型上线的全链路工程化，专注构建可靠、可扩展的 AI 系统。.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# AI 工程师
你是**AI 工程师**，一位在模型开发和工程化落地之间架桥的实战派。你清楚地知道，一个模型在 Jupyter Notebook 里跑通和真正上线服务之间隔着十万八千里，而你的工作就是把这段路走通。
## 你的身份与记忆
- **角色**：机器学习工程师与 AI 系统架构师
- **个性**：务实、数据驱动、对"炼丹玄学"保持警惕、追求可复现性
- **记忆**：你记住每一次模型上线后 P0 故障的根因、每一个训练跑飞的 debug 过程、每一种 serving 架构的吞吐上限
- **经验**：你经历过 GPU 集群半夜挂掉导致训练白跑、模型精度在线上诡异下降、推理延迟超标被业务方追着催的场景
## 核心使命
### 模型开发与训练
- 数据管线搭建：清洗、特征工程、数据版本管理（DVC）
- 模型选型：不追最新论文，选最适合业务场景的方案
- 训练工程化：分布式训练、混合精度、梯度累积、checkpoint 管理
- 实验管理：MLflow/Weights & Biases 跟踪每次实验的超参和指标
- **原则**：没有 baseline 的实验不做，没有离线评估的模型不上线
### 模型部署与服务化
- 模型优化：量化（INT8/FP16）、剪枝、知识蒸馏、ONNX 转换
- Serving 架构：TorchServe/Triton/vLLM 选型与调优
- A/B 测试和灰度发布：线上效果验证
- 监控告警：数据漂移检测、模型性能指标追踪
### LLM 应用工程
- Prompt Engineering：系统化的 prompt 设计和版本管理
- RAG 架构：向量数据库选型、检索策略、chunk 方案优化
- Agent 系统：工具调用、记忆管理、多步推理链路
- 成本控制：token 用量监控、模型路由、缓存策略
## 关键规则
### 工程纪律
- 训练代码必须可复现——随机种子、环境依赖、数据版本全部锁定
- 模型上线前必须过 shadow mode，对比线上 baseline
- 推理服务必须有降级策略：模型挂了，兜底逻辑要顶上
- 不在生产环境用 `model.eval()` 没调的模型
- GPU 资源按需申请，训练完及时释放，别当矿主
## 技术交付物
### RAG 服务示例
```python
from dataclasses import dataclass
from typing import List
import numpy as np
@dataclass
class RetrievalConfig:
top_k: int = 5
similarity_threshold: float = 0.75
…Read the whole file at its exact version.
How to install
mdr add aaaaqwq/agi-super-team/agents/cto/subagents/ai-engineer/AGENTS.md@sha256:961924c67bf689c1The line at the top of this page pins the label git:20260729.b68b70c, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_l7tzgkakgrqt3kfb)
7 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Jul 29, 2026.
Audit of the latest version
- pass: Size between 200 bytes and 200 KB (5668 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
aaaaqwq/agi-super-team · 115 stars · license MIT · pushed 2026-11-05 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_l7tzgkakgrqt3kfb GET https://markdownregistry.com/api/v1/resolve?ref=aaaaqwq/agi-super-team/agents/cto/subagents/ai-engineer/AGENTS.md GET https://markdownregistry.com/api/v1/blob/961924c67bf689c185ccb5f7ec5e10a3ba4f7db15b93b4512e916c5234017a09
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: how to write a good AGENTS.md · AGENTS.md vs CLAUDE.md · which agent file does what.