AI Product Manager candidate · AI-native systems · Education AI · Cloud and Agent workflows

Email · GitHub · Tencent Cloud CSIG · AI Product

I am a master's student in Big Data Technology and Engineering, building toward AI product roles that sit between product judgment, model evaluation, data systems, and engineering delivery.

我的主线不是“会调用模型”，而是把 AI 能力做成能被业务使用、被指标验证、被工程稳定承载的产品闭环。

real scenario
  -> product problem
  -> AI capability boundary
  -> data / evaluation system
  -> usable workflow
  -> measurable iteration

• Tencent Cloud CSIG: AI Agent platform documentation, public capability evidence, cloud product information architecture, and automated documentation governance.
• Master's thesis work: an education AI loop with three layers: controllable synthetic data, Result/Process benchmark evaluation, and a teaching workspace.
• AI product job search: Qingdao-oriented AI product, AI platform, product technology, and solution roles.

• Education AI Benchmark and Decision System Built a benchmark framing that separates answer correctness from teaching-process quality, supporting model selection with leaderboard, bias analysis, and second-pass calibration.

• Controllable Synthetic Data for Education AI Designed CTSP-style constrained task generation around “what to teach” and “how to teach,” aiming to make synthetic data more steerable, auditable, and useful for downstream training.

• WenYunZhiTu Teaching Workspace A student-teacher-admin AI teaching workflow: student questions, projectized learning traces, challenge confirmation, teacher verification, and SFT/DPO sample export.

• Tencent Cloud TCCLI Skill Standardizes Tencent Cloud API operations through tccli, making cloud workflows more agent-operable. Repository

• Agent Doc Architect Agent-friendly documentation architecture: task-first information architecture, runbook loops, output conventions, and reusable templates. Repository

• TKE CLI Guide CLI-first guide materials for Tencent Kubernetes Engine workflows. Repository

• Image2PPT Converts courseware screenshots into editable PPT through image layering, super-resolution, OCR, and VLM reasoning. Repository

• AI Resume Generator Explores structured resume generation, content enhancement, and JD-aware rewriting for AI-assisted job search workflows. Repository

• QA from PDFs Builds validated QA JSONL from paired question and answer PDFs for education datasets and benchmark construction. Repository

Product strategy · AI evaluation · synthetic data · education AI · cloud product · agent workflow · Python · TypeScript · Shell · Next.js · Supabase

The profile avoids third-party stats cards so it renders reliably in recruiter, mobile, and low-connectivity views. For a quick read, start here:

• Product and AI workflow proof: Image2PPT, AI Resume Generator
• Cloud and agent workflow proof: Tencent Cloud TCCLI Skill, TKE CLI Guide, Agent Doc Architect
• Education data proof: QA from PDFs

I am open to AI product roles where product judgment, AI evaluation, data infrastructure, cloud delivery, and industry landing all matter.

• Email: kerwin01130224@gmail.com
• GitHub: Kerwin0224

_{Profile README maintained in Kerwin0224/Kerwin0224.}