Skip to content
Kevin拾玖秋— home
中文

About — Kevin / 拾玖秋

From Architecture
to AI Engineering.

I'm a developer focused on AI applications and agent engineering. I build AI systems around real workflows — agents, retrieval, tools, evaluation and delivery.

01Path

Where I am now

  1. 2023Started exploring Generative AI & LLMs.
  2. 2024AI application development & training.
  3. 2025AI development in production-oriented scenarios.
  4. 2026Agent Engineering / RAG Evaluation / AI Coding / Open Source.

I also build in public and contribute fixes back to open-source projects.Open-source work

Current focus

  • AI Agent
  • RAG
  • LLM Workflow
  • Enterprise Knowledge Base
  • AI Tool Calling
  • Local-first AI
  • AI Coding Agent
  • AI Product Engineering

02Engineering principles

How I Build AI Products

  1. 01. Problem Before Model

    Solve the business problem first. Choose the model after.

  2. 02. Workflow Before Agent

    If a workflow can solve it, it doesn't need an agent. Autonomy is a cost, not a feature.

  3. 03. Evaluate Retrieval

    RAG is not a vector database with a prompt on top. Retrieval has to be measured.

  4. 04. Evidence Matters

    Without enough evidence, the system should not pretend to know the answer.

  5. 05. Demo Is Only the Beginning

    Production AI also needs:

    • State
    • Evaluation
    • Testing
    • Error Handling
    • Privacy
    • UI
    • Delivery

03The full chain

More than a model call.

My focus is not only the model API call. It is the whole application chain — from the user's workflow to evaluation, UI and delivery.

Select a layer to see what it means and where it shows up.

01 / User

Start from real users and their workflow, not from what the model can do.

04Background

Architecture taught me how to structure complex problems.

My undergraduate degree is in Architecture. AI engineering gave me a new medium to build solutions.

What design training left me with

  • 01Complex constraints
  • 02System thinking
  • 03Information hierarchy
  • 04Visual communication
  • 05Design iteration

Architecture shows up in how I structure systems. The work itself is AI engineering.

05Secondary practice

Architecture × AI

A continuing line of work on how AI fits into the architectural design process — from analysis diagrams to design text.

  1. 01AI Architecture Diagram
  2. 02AI Concept Analysis
  3. 03Architecture Visualization
  4. 04ComfyUI Workflow
  5. 05AI Design Text
  6. 06CAD / SketchUp / AI Workflow
Architecture AI training
Abstract isometric massing studyAAMASSING / 01SCALE — ABSTRACTAI-ASSISTED
Abstract massing study — generated diagram, not a project image.

06Tech stack

Tools I work with

AI Application

  • LLM Application
  • AI Agent
  • RAG
  • Knowledge Base
  • Prompt Engineering
  • Tool Calling
  • Workflow Design
  • Human-in-the-loop
  • Retrieval Evaluation

Frameworks

  • Dify
  • Coze
  • LangChain
  • smolagents

Model / API

  • DeepSeek
  • Qwen
  • DashScope
  • SiliconFlow

Retrieval

  • FAISS
  • BM25
  • BAAI / bge-m3
  • text-embedding-v3

Engineering

  • Python
  • Git
  • pytest
  • ruff
  • REST API
  • JSON / Structured Output

AI Coding

  • Claude Code
  • OpenAI Codex