TongYi (HK) Limited · 桐燚科技有限公司

Who We Are 我們是誰

Hong Kong AI Hub

Focused on AI Education & Enterprise AI Applications 專注於人工智能教育與企業人工智能應用

Who We Are我們是誰

TongYi (HK) Limited is a Hong Kong-incorporated AI company founded by Cai Jian'en (Elvis). Leveraging next-generation information technologies such as Artificial Intelligence, Big Data, and Cloud Computing, we empower brands and merchants to achieve sustainable growth and long-term market traction. 桐燚科技有限公司是一家於香港註冊成立的人工智能公司，由蔡建恩（Elvis）創立。憑藉人工智能、大數據及雲計算等新一代資訊技術，我們助力品牌與商家實現可持續增長及長期市場影響力。

At our core, we specialise in AI education and training, and enterprise AI system deployment — building verifiable, compliant AI capability layers for real business use cases. 我們的核心業務專注於人工智能教育培訓，以及企業人工智能系統部署——為真實商業應用場景建立可驗證、合規的人工智能能力層。

What We Do | Two Core Service Pillars 我們的業務 | 兩大核心服務支柱

Two Core Service Pillars

AI Professional Training人工智能專業培訓

Practitioner-focused AI courses for adults: AI Agent development, Python computer vision, AIGC video generation, 3D spatial modelling & AI data annotation. 面向成人的實務型人工智能課程：AI智能體開發、Python電腦視覺、AIGC影片生成、3D空間建模與人工智能數據標註。

Enterprise AI Systems & Consulting企業人工智能系統與諮詢

Industrial vision, logistics scheduling, RAG knowledge bases, RPA automation (500+ app integrations), green compliance data (MRV/CBAM); MoCap data pipeline processing & LoRA/video model fine-tuning, powering scalable AIGC short drama, brand content and creative IP production. 工業視覺、物流調度、RAG知識庫、RPA自動化（500+應用整合）、綠色合規數據（MRV/CBAM）；動作捕捉數據管道處理與LoRA/影片模型微調，賦能可規模化的AIGC短劇、品牌內容及創意IP製作。

AI Matrix Popular Product Drainage System人工智能矩陣爆款引流系統

Multi-platform management, mix and cut high-quality videos in batches, batch and scheduled distribution, customer acquisition marketing code, private domain applet, private messages, collection of comments. 多平台管理、批量混剪優質影片、批量定時分發、獲客營銷代碼、私域小程序、私信與評論收集。

How to Integrate into the Institution's Existing Curriculum 怎麼排進機構現有的課程

Cooperation Boundaries: Who Does What 合作邊界：誰做什麼

We Have Worked With Training Institutions 我們和培訓機構打過交道

Selected Work | Enterprise AI 精選案例 | 企業人工智能

MoCap Intelligence · LoRA Fine-Tuning · AIGC 動作捕捉智能 · LoRA微調 · AIGC

AI Motion Capture人工智能動作捕捉

Convert motion capture sequences into structured training datasets for downstream AI models — posture classification, anomaly detection, fatigue monitoring on factory floors. 將動作捕捉序列轉換為結構化訓練數據集，供下游人工智能模型使用——姿態分類、異常檢測、工廠現場疲勞監測。
Fine-tune video generation models (LoRA) with industrial motion data — producing AIGC content for training simulations, safety briefings, and SOP documentation. 利用工業動作數據微調影片生成模型（LoRA）——製作用於培訓模擬、安全簡報及標準作業程序（SOP）文件的AIGC內容。
Build end-to-end AIGC production pipelines for industrial clients — AI generated instructional videos, compliance demonstrations, and branded content at a fraction of traditional production cost. 為工業客戶構建端到端AIGC製作管道——以傳統製作成本的一小部分，生成人工智能教學影片、合規演示及品牌內容。

$AI Motion Capture stills$

Above: 《染殺》, an AI-generated short film produced for short film platform with 佛山炎宸 motion capture AI stack 上圖：《染殺》，一部由佛山炎宸動作捕捉AI技術棧為短劇平台製作的人工智能生成短片

POSTURE RIDER — AI Checks Your Form 姿勢騎士 — 人工智能檢查您的姿勢

Computer Vision · AI Labelling · Industrial Training Pipeline 電腦視覺 · 人工智能標註 · 工業訓練管道

AI Vision & Data Capture人工智能視覺與數據採集

Deploy camera-based AI vision on sites to capture real-time operational data — worker movement, equipment status, load patterns. 在現場部署基於攝影機的人工智能視覺系統，實時採集運營數據——工人動態、設備狀態、負載模式。
Structure captured footage into labelled datasets using CVAT-based annotation workflows — bounding boxes, pose keypoints, segmentation masks built for YOLOv8 training. 利用基於CVAT的標註工作流程，將採集畫面結構化為標註數據集——為YOLOv8訓練構建邊界框、姿態關鍵點及分割遮罩。
Train and fine-tune custom vision models on client-specific industrial scenes — enabling automatic defect detection, safety compliance monitoring, and throughput analytics. 針對客戶特定工業場景訓練及微調定制視覺模型——實現自動缺陷檢測、安全合規監測及產能分析。
Deploy trained models back into live camera feeds via edge inference — closing the full loop from raw capture to actionable AI decision layer. 通過邊緣推理將訓練完成的模型部署回實時攝影機畫面——閉合從原始採集到可執行人工智能決策層的完整循環。

CAES | AI Film and Television Industrial Production CAES | 人工智能影視工業化製作

AI Video Production Case: Chinese Comic Style人工智能影片製作案例：中國漫畫風格

For the first time in AI comics, Chinese-style visuals and Disney-style dynamic performances are achieved. 首次在人工智能漫畫中，實現中國風視覺與迪士尼式動態表演的結合。
Relying on the industrialized AIGC process, it restores character acting skills and maintains visual consistency on par with benchmark theatrical animated films, while greatly shortening the production cycle. 依託工業化AIGC流程，還原角色表演技巧，並保持與標杆院線動畫電影相當的視覺一致性，同時大幅縮短製作週期。
Powered by our self-developed AI dynamic control tech, we maintain visual consistency through intense action and subtle emotional beats alike — eliminating the stiffness, flicker, and continuity errors often seen in legacy AI video models. 憑藉我們自主研發的人工智能動態控制技術，我們在激烈動作與細膩情感節拍中均保持視覺一致性——消除傳統人工智能影片模型常見的僵硬、閃爍及連貫性錯誤。

TY TongYi (HK) Limited