ABOUT

Yongliang Zhang

Kael · AI Technical Director

AI Technical Director · 17 Years · 20+ Products Shipped

17Years building software
20+Products 0→1 shipped
70Peak team size
100k+Users on core product
16+AI certifications

I build AI that pays for itself.

For 17 years I have been shipping software — from Java engineer to technical director across healthcare, legal-tech, traffic big-data and new media. Today I lead AI at a hospital group in Tianjin: internet-hospital products, a medical AI platform, RAG knowledge bases, ChatBI dashboards, and data governance that turns messy HIS records into AI-ready assets.

What I actually do: take LLMs, RAG and agents — and convert them into products with measurable outcomes. 100k+ registered users, 386k annual active visits, ¥3.25M angel funding contributed. I care about the version that ships, not the demo that impresses.

What I Do Now

At a hospital group in Tianjin I run the AI platform: intelligent triage, four-quadrant customer analysis, AI workbench for doctors, medical data cleaning, Dify agents, Spring AI gateway, and organization-wide AI tooling that made our own teams faster.

My Role: Forward Deployed Engineer

The FDE model — born at Palantir, now standard at AI companies — puts engineers inside the customer’s operation: translate fuzzy business pain into shipped AI solutions, and own the outcome. That is literally what I have been doing in hospital corridors for years. Technical Director is my title; FDE is how I work — half on the business floor, half writing the solution.

How I Work

01

Ship, not demo

The version that goes live beats the demo that impresses. Every AI project ends with a metric that moved.

02

Facts first

Everything I publish is checked against primary sources before it goes out. Numbers get timestamps.

03

Half in the field

Requirements live in the workflow, not the requirement doc. I sit where the work happens.

04

Build once, earn forever

Products should outlive projects. I design for the second year, not the launch week.

What I Work With

LLM appsRAGAgentsPrompt engineeringDifySpring AIMilvusChatBI / SupersetData governanceJava/SpringFull-cycle deliveryAI-assisted R&D (Codex / Claude / OpenCode)

Selected Certifications

Google Prompting Essentials · IBM AI Product Manager · IBM GenAI (UI/UX, Java) · Microsoft AI & ML Engineering · NVIDIA AI Foundations · iFLYTEK Prompt/RAG/Agent Engineer · Alibaba AI Trainer (Adv.) · plus 15+ more · Speaker-stage participant, 2023 CCHIO cancer-prevention science-popularization initiative (with CACA, Xinhua, and hospital experts)

About This Site

This is where I write things down: teardowns of AI frameworks, honest talks about the AI era, and quick takes on trending news. Five languages, one author.

B.S. Computer Science, Nankai University

Frequently Asked Questions

Who is Kael Zhang?

Kael Zhang (永亮) is an AI technical director with 17 years of software engineering experience. He currently leads AI at a hospital group in Tianjin, and is the author of three AI columns — Shiwen Talks, Tech Teardown and Hot Topics — published daily in five languages.

What does he actually do?

He turns large language models, RAG systems and AI agents into shipped software products with measurable outcomes: 100k+ registered users, 386k annual active visits, and a ¥3.25M angel round contributed across projects.

Where can I follow his work?

Chinese content: WeChat official account 「永亮说AI」. English content: Substack (kaeltalksai.substack.com). Overseas social: X @jrrexliang, YouTube @kaelzhang, GitHub kaelzhang321.

How do I contact him?

Via DM on X (@jrrexliang) or the contact links in the site footer. He does not take outsourcing projects — only AI-product and content-related collaboration.