INTERVIEW INTRODUCTION · 05 MIN

冉熙Xi Ran 中国科学院大学 · 2027 届计算机技术硕士UCAS · M.S. Computer Technology, 2027

从业务 Agent 编排到云原生 Runtime，把模型能力做成可验证的系统。

From Agent orchestration to cloud-native runtimes, I turn model capabilities into verifiable systems.

University of Chinese Academy of Sciences

国科大杭高院 × 中科院计算所 × 浙江大学联合培养 Joint training across HIAS, ICT and Zhejiang University

Agent RuntimeSandbox · Warm Pool · Session

Agent OrchestrationTool Use · Routing · Recovery

工程验证Engineering ProofBenchmark · E2E · Open Source

冉熙在电脑前工作 — HANGZHOU · 2026AI AGENT ENGINEERING

面试官您好，我叫冉熙，是中国科学院大学计算机技术专业 2027 届硕士。我的定位是 Agent 系统工程：在 Bybit，我把专家排查拆成可编排、可恢复的 Agent 工具链；在华为，我进一步做到 Sandbox Runtime、性能验证和云原生上游交付。

Hello, I am Xi Ran, a 2027 master's candidate in Computer Technology at UCAS. I focus on Agent systems engineering: orchestrating recoverable toolchains for risk decisions at Bybit, then moving deeper into sandbox runtimes, performance validation and cloud-native upstream delivery at Huawei.

01 / TRAJECTORY

四段经历，一条 Agent 系统主线Four roles, one Agent systems trajectory

02

从应用交付，到 Agent 决策，再到 Agent Runtime：
负责的系统边界持续向下延伸。

From application delivery to Agent decisions and Agent runtimes, my ownership has moved deeper into the stack.

主线：业务问题 → Agent 编排 → Runtime 与上游Path: business problem → orchestration → runtime and upstream

2023.10–2024.06

趣链科技 · 企业应用研发Hyperchain · Enterprise Applications从业务需求到全栈交付Requirements to full-stack delivery

2025.06–2026.06

浙江大学 · 量化 Agent 系统Zhejiang University · Quant Agents多 Agent 决策与全栈平台Multi-Agent decisions and platform

2025.12–2026.06

Bybit · 风控 Agent 算法Bybit · Risk Agent Algorithms12 个 Tool 编排专家排查12 tools orchestrating expert analysis

2026.06–至今

华为 · Agent InfraHuawei · Agent InfraSandbox Runtime、Benchmark 与上游Sandbox runtime, benchmarks and upstream

10×Bybit Agent 提效Bybit Agent speedup

<1sSandbox 本地 p50local Sandbox p50

20华为实习相关上游合入 PRHuawei internship upstream PRs

我的主线不是换行业，而是逐步扩大 Agent 系统的责任边界：先交付完整应用，再负责 Agent 的工具编排和决策质量，最后深入 Sandbox Runtime、性能与云原生上游。下面我用两个 STAR 案例说明。

The trajectory is an expanding Agent systems boundary: complete applications, then tool orchestration and decision quality, and now sandbox runtimes, performance and cloud-native upstream work. The next two slides show concrete STAR cases.

02 / CASE STUDY · BYBIT

从 40 分钟人工流程，到 4 分钟全市场 AgentFrom a 40-minute manual flow to a 4-minute market-wide Agent

03

S / T：专家需要串行完成多轮取数、关联和验证，单币种耗时 20–40 分钟；我的任务是在查询资源受限的条件下实现可批量、可复核的 Agent 排查。

S / T: Experts spent 20–40 minutes per symbol on sequential retrieval, correlation and verification. I owned a batchable, reviewable Agent workflow under constrained query resources.

4 MIN端到端扫描全市场end-to-end, market-wide

10×人工流程提效workflow speedup

0→4同轮验证异常检出anomalies in one validation

Action：不是让模型直接给结论，而是把专家判断拆成有输入、阈值、失败处理和证据输出的工具链。

Action: The model never guessed the verdict directly; expert reasoning became tools with inputs, thresholds, failure handling and evidence.

01取数Retrieve6 SQL · Hive/Presto

02编排Orchestrate12 Tools · 5-way

03判定Decide交易多样性 · 8维评分diversity · 8D score

04恢复RecoverCDP · Checkpoint

05证据Evidence可解释报告reviewable report

约束：Constraint: 40GB 查询限制下用显式 UID、短路和分片规避 OOMexplicit UIDs, short-circuiting and sharding under a 40GB query limit

Result：Result: 累计识别 10+ 异常团伙，结论保留可复核证据10+ abnormal groups identified with reviewable evidence

背景是异常账户分散，专家要串行做多轮取数、聚类和行为验证，单币种需要二十到四十分钟，无法批量扩展。我的任务是在 Hive/Presto 查询资源受限的情况下，把这套判断变成可自动执行、失败可恢复、结论可复核的 Agent。我从零设计五阶段 Pipeline，开发六条协作 SQL 和十二个 Tool，使用五路并行与条件路由；算法上提出交易多样性信号并构建八维评分，工程上用显式 UID、短路、分片、CDP 自动恢复和断点续传处理 OOM 与长任务失败。最终把流程压缩到四分钟全市场扫描，实现十倍提效；同轮验证从零个异常提升到四个，累计识别十多个异常团伙。

Abnormal accounts were distributed, and experts spent 20–40 minutes per symbol on sequential retrieval, clustering and behavioral verification. I owned an automated, recoverable and reviewable Agent workflow under constrained Hive/Presto resources. I designed a five-stage pipeline with six SQL queries and twelve tools, using five-way parallelism and conditional routing. I added trade diversity and an eight-dimension score, plus explicit UID queries, short-circuiting, sharding, CDP recovery and checkpoints. The result was a four-minute market-wide scan, a 10x speedup, 0-to-4 anomalies in one validation, and more than ten groups identified cumulatively.

03 / AGENT INFRA · HUAWEI

华为实习｜Agent Infra 与云原生控制面研发Huawei Internship | Agent Infra & Cloud-Native Control Planes

04

AgentCube × Karmada · 2026.06–09 · 从架构设计、性能验证推进到上游交付AgentCube × Karmada · Jun–Sep 2026 · Architecture, performance validation and upstream delivery

AgentCube Agent Sandbox 基础设施Agent Sandbox infrastructure

K8s Control Plane Session · API · Router · Store

Node Runtime Runtime · Provider · Warm Pool

并发性能Concurrency建立 create–run–delete Benchmark，定位 Warm Pool 容量瓶颈Built a create-run-delete benchmark to locate Warm Pool capacity limits
架构设计Architecture参与 K8s / Node 双层控制面设计，解耦资源编排与高频生命周期Co-designed K8s and Node control planes to separate orchestration from fast lifecycle operations
生命周期兼容Lifecycle完成 v0.4.6 创建、接管、执行、删除与回补链路适配Adapted v0.4.6 create, adopt, run, delete and pool-refill flows
CI 性能优化CI Performance使用原生 builder 与 Go 交叉编译优化多架构构建Optimized multi-architecture builds with native builders and Go cross-compilation

Karmada Kubernetes 多集群编排Kubernetes multi-cluster orchestration

Control Plane 调度器 / 控制器Scheduler / Controllers

Delivery Work

Execution 成员集群Member Clusters

多组件调度Scheduling完成结果持久化、扩缩容触发与失败保护Added result persistence, scaling triggers and failure protection
身份恢复Recovery实现叶子证书轮换，保留 CA、SAN 与 external-etcd 配置Implemented leaf certificate rotation while preserving CA, SAN and external-etcd settings
控制器可靠性Reliability修复资源清理、事件触发和测试隔离问题Fixed resource cleanup, event triggering and test isolation issues

7项核心工作core engineering tasks

20个实习相关上游 PR 合入internship-related upstream PRs merged

END-TO-END方案设计 → 代码实现 → 实验验证 → 上游交付design → implementation → validation → upstream delivery

华为三个月实习的工作集中在 AgentCube 和 Karmada 两个上游项目。AgentCube 侧，我建立 create、run、delete 分阶段并发 Benchmark，参与 K8s 集群控制面与 Node 节点控制面的双层架构设计，并完成 v0.4.6 生命周期兼容和 CI 多架构构建性能优化。Karmada 侧，我围绕多组件调度完成结果持久化、扩缩容触发与失败保护，并实现叶子证书轮换、修复控制器清理和事件触发问题。三个最强结果是：组件级调度结果持久化随 Karmada v1.19 正式发布；本地 k3s 十并发下，warmPoolSize 从二调整到十后，Sandbox p50 从 7.315 秒降至 436 和 565 毫秒；多架构完整构建从 1610 秒降至 331 秒，减少 79.4%，同时保留 amd64 和 arm64 产物。整个实习形成了从方案设计、代码实现、实验验证到上游交付的完整研发经验。

My three-month Huawei internship focused on two upstream projects, AgentCube and Karmada. In AgentCube, I built a phased create-run-delete concurrency benchmark, co-designed the K8s and Node control-plane split, and delivered v0.4.6 lifecycle compatibility plus CI multi-architecture build performance improvements. In Karmada, I worked on multi-component scheduling result persistence, scaling triggers and failure protection, then implemented leaf certificate rotation and controller reliability fixes. The strongest evidence is component-level scheduling persistence released in Karmada v1.19, Sandbox p50 reduced from 7.315 seconds to 436 and 565 milliseconds on local k3s at concurrency ten when warmPoolSize changed from two to ten, and a multi-architecture build reduced from 1610 to 331 seconds while preserving amd64 and arm64 outputs.

APPENDIX / OPEN-SOURCE PORTFOLIO

21 个上游仓库，按能力域完整展开21 upstream repositories, organized by capability

A1

21上游仓库repositories

70已合入 PRmerged PRs

21协作中 PRPRs in review

绿色：已合入 · 橙色：协作中Green: merged · Orange: in reviewGitHub 状态截至 2026.10.07Observed October 7, 2026

这页展示完整开源贡献版图，不计本人 fork、自有仓库或仅用于报名的 PR。21 个上游仓库按 Agent 与上下文、Runtime 与 Kubernetes、工具与社区三组展示；截至 2026 年 10 月 7 日，共 70 个 PR 已合入、21 个 PR 协作中。绿色只表示 GitHub merged 状态，橙色项目仍由 maintainer 决定后续是否合入。本页不是华为实习成果统计；华为实习范围仍以 AgentCube、Karmada 与 Kubernetes SIGs Work API 的 21 个已合入 PR 为准。

This appendix shows the full open-source portfolio, excluding personal forks, owned repositories, and registration-only PRs. Twenty-one upstream repositories are grouped into Agents and Context, Runtime and Kubernetes, and Tools and Community. As of October 7, 2026, 70 PRs are merged and 21 remain in review. Green indicates GitHub merged status only; orange items still depend on maintainer decisions. This is not the Huawei internship outcome count, which remains 21 merged PRs across AgentCube, Karmada, and Kubernetes SIGs Work API.