Professional Resume

Factory AI Adoption: The Data Foundation Is the First Barrier — Using AI Itself to Tackle Legacy Equipment Protocol Challenges

Professional Article | Published 2026-10-15

Sole author and publisher (personal viewpoint article)

Diagnoses why so many manufacturing AI projects stall mid-way: the real bottleneck is not the algorithm but the collection of high-quality underlying data. The article argues for the correct sequence — build the data foundation first, then deploy upper-layer AI — and proposes a strategy of AI combined with self-developed edge computing boards for data reverse engineering: on-site packet capture, AI packet analysis and feature extraction, and automatic generation of protocol layers, so both legacy proprietary protocols and modern standards can be integrated without vendor documentation or mass machine replacement.

Read the full article · 繁體中文版

Plant Manager

Changshu Jingyuan Computer Co., Ltd.
2024-09 — 2026-08 · Changshu / Suzhou, China

Led production, quality, engineering, customer service, EHS, HR and ESG to improve operating efficiency and profitability.

• Used AI to analyze process parameters and optimize manufacturing, delivering RMB 2M+ in annual cost benefits.
• Built a private enterprise AI agent for regulatory knowledge, equipment maintenance, quality and management decisions. A 30% reduction in quality staffing requirements is an estimated benefit.
• Advanced lean manufacturing through OEE, downtime and changeover management, with daily and weekly production reviews.
• Led implementation, maintenance and audit preparation for ISO 9001, IATF 16949, ISO 14001 / 45001, QC 080000, TISAX, RBA and EcoVadis systems.
• Served as SHE Manager with overall production safety responsibility, overseeing EHS risk controls, emergency drills and leadership development.
• Directed engineering and intellectual property strategy, introducing an Engineering AI Agent to review patent gaps, competing technologies and design-around risks.
• Participated in AI research on keycap pull-force prediction using material, opening dimensions and force data to explore design optimization and offline validation.

Management Director

Shengwang Auto Parts (Kunshan) Co., Ltd.
2022-01 — 2024-09 · Kunshan, China

Integrated HR, EHS, administrative costs, regulatory compliance, government relations, quality systems and automation planning.

• Applied PDCA to workforce assessment, quality improvement and cost control; supported retention and long-term incentives for key talent.
• Planned PLC / Kepware / OPC communication standards to connect equipment data with management systems.
• Applied APQP, SPC, FMEA, MSA, PPAP and 8D to automotive and electronics quality management.
• Established and led an innovation center to advance patent applications and proprietary technology.

Co-Founder / Executive Vice President

Suzhou Bochengkang Technology Co., Ltd.
2013-01 — 2021-12 · Suzhou, China

Co-founded the business and built its operations from the ground up, spanning product positioning, R&D, teams, supply chain, manufacturing and business development.

• Led wireless control, cloud IoT and hardware–software integration for lithium-battery e-bikes. Related products received a German iF Design Award and patents in multiple countries.
• Developed e-bike communication protocols achieving over 65% hardware compatibility, with OTA update support.
• Helped establish SMT manufacturing, quality and automation systems; cross-functional improvements reduced material inventory by more than 10%.
• Built a sales team and FABE sales SOP, using exhibitions, customer development and cross-industry partnerships to expand markets.
• Planned E-bike App and administration architecture, studying vehicle diagnostics, e-commerce, after-sales service, communities and IoT integration.

Quality Assurance Manager

IEI Integration Corp.
2010-08 — 2012-12 · Shanghai, China

Managed industrial PC validation, quality systems, customer quality services and the BIOS R&D team.

• Oversaw electrical, software, environmental, vibration and EMC testing, laboratory resources, NPI, DCC and internal / external audits.
• Led an interactive product-catalog app integrating product, inventory and supplier information to reduce printing costs and support mobile sales.

Deputy Supervisor, Product Validation

Inventec
2004-07 — 2010-08 · Shanghai, China

Managed quality validation and customer communication for HP notebook computers.

• Coordinated software, hardware and mechanical test plans, quality audits, issue tracking and retesting to align quality requirements with project schedules.

Software / Firmware Test Engineer

Tatung
2003-04 — 2004-06 · Taipei, Taiwan

Worked on system analysis, OEM OS build automation and WHQL testing.

• Tested and modified systems to customer requirements, building practical experience in network applications and databases.

Professional Article | The Data Foundation of Factory AI Adoption

Professional Article | Published 2026-10-15

Sole author and publisher (personal viewpoint article)

Diagnoses why so many manufacturing AI projects stall mid-way: the real bottleneck is not the algorithm but the collection of high-quality underlying data. The article argues for the correct sequence — build the data foundation first, then deploy upper-layer AI — and proposes a strategy of AI combined with self-developed edge computing boards for data reverse engineering: on-site packet capture, AI-driven packet analysis and feature extraction, and automatic generation of protocol layers, so both legacy proprietary protocols and modern standards can be integrated without vendor documentation or mass machine replacement.

• Core thesis: the data and protocol barrier predates AI — it already plagued factories during Industry 4.0 digitalization; AI performance depends entirely on input quality.
• Industry critique: most software vendors sell only upper-layer AI tools — "selling a magic wand while charging customers to wield it" — with no data collection or protocol parsing underneath, even cutting-edge Agents cannot function.
• Proposed solution: AI plus self-developed edge boards for data reverse engineering, compressing protocol reverse-engineering from months of manual engineer work into an automated pipeline.
• Conclusion: skipping the data foundation looks like a shortcut, but the hidden risk is a project that becomes a demo unable to run stably on a real production line.

Read the research and source notes

AI Surface Defect Classification & Explainability

Applied AI Vision | 2026

Data pipeline, model training, explainability analysis and demo interface — independently delivered

A reproducible vision-classification build on an industrial surface-defect problem: constructed a data pipeline from a public industrial dataset, compared CNN and Transformer transfer learning under an identical protocol, validated model evidence against human annotations using Grad-CAM, and packaged the result into a live inference demo. The goal was to test whether "accurate is not enough — a line AI must explain its call" is workable in practice.

• Dataset: NEU-DET steel surface defect benchmark (shop-floor collected, 1,770 images x 6 classes: crazing, inclusion, patches, pitted surface, rolled-in scale, scratches), stratified train/val/test split with a fixed seed (1,236/264/270), verified leak-free.
• Models: ResNet18 (CNN) and ViT-B/16 (Transformer), both ImageNet-pretrained transfer learning; both reached 100% accuracy and macro-F1 1.00 on a 270-image held-out test set.
• Explainability: Grad-CAM heatmaps overlaid against ground-truth boxes, 12/12 aligned; observed CNN heatmaps concentrating on the defect body while ViT attention spread wider.
• Engineering: NVIDIA RTX 5060 (Blackwell) + PyTorch CUDA 12.8 + bf16 mixed precision; fully scripted and reproducible.
• Deliverables: data pipeline, training/evaluation scripts, confusion matrix and classification report, Grad-CAM comparison, Gradio inference demo.

Read the research and source notes

Wafer-to-Market Semiconductor Back-End Experience

Semiconductor Supply Chain & Commercialization Project | 2021

Funding and supply chain integration, test and packaging coordination, and sales

During the 2021 pandemic and chip shortage, sourced 16 SXX wafers through industry channels, coordinated funding and supply chain resources, and completed the workflow from wafer sourcing through testing, packaging and final verification to sales.

• Conducted pre-packaging verification, assessed product test status (second/third test rounds), independently sourced Probe/CP Test providers, and coordinated wafer probing and yield assessment.
• Engaged with and visited multiple packaging and testing facilities in mainland China; selected packaging partners, confirmed packaging parameters, and coordinated packaging completion and final product verification.
• Completed the Wafer → Probe/CP → Yield Assessment → Package Engineering → OSAT → Final Verification → Sales workflow, building practical experience in semiconductor back-end supply chain integration and commercialization.

Read the research and source notes

LINE Scan-to-Order SaaS (PickMy Order)

Independent Build | 2026

Product definition, system architecture, and development lead

A LINE scan-to-order and light-membership SaaS for offline stores, built around scan-to-order, loyalty points, and prepaid service cards — driven independently from product definition and full-stack development through automated testing (described generically; no project-specific details).

• Light membership system: scan-to-order + points + prepaid service cards closed loop, with 197 automated regression tests; every release fully verified before deployment.
• Production scheduling optimization: defined customer wait time, one-trip fulfillment rate, and capacity utilization as the objective function, and built a re-runnable evaluation harness (fixed random seed) to quantitatively compare candidate algorithms — reproducible and regression-safe conclusions.
• Manufacturing methodology transferred: applied daily/weekly review cadence and quality-engineering thinking from mass production to SaaS iteration and requirements management.

Read the research and source notes

0→1 Digital Platform Build & Launch

Independent Build | 2026

End-to-end owner: architecture, development, testing, production launch

Took a multi-role service platform from architecture design and full-stack development through automated testing to production deployment on cloud infrastructure — completed end to end in personal time (described generically; no project-specific details).

• Five-endpoint system architecture: customer / admin / vendor / partner portal / public website, with multi-role access control and security boundaries designed up front.
• Production-grade deployment & operations: Linux server provisioning, Nginx reverse proxy, systemd services (auto-start, auto-recovery), and a one-command deployment script — code-to-public-web in under 15 minutes.
• Full compliance chain: domain registration, DNS, ICP filing, HTTPS certificates, plus WeChat mini-program registration/certification and payment ecosystem integration.
• Quality engineering: 80+ automated regression tests plus targeted test suites gating every release; automated screenshot tooling produced a 60+ page system manual.
• Complex business modeling: state-machine workflows, funds-flow and reconciliation logic, exception compensation and risk-control thresholds — compliance rules built into system design.

Read the research and source notes

AI Process & Engineering Parameter Optimization

Workplace Project Research | Changshu Jingyuan

Participated in engineering pull-force prediction and supported manufacturing AI applications through plant management.

Translated manufacturing and design problems into measurable parameters, prediction targets and validation workflows. Studied key pressing, paint fixtures, pogo-pin force, keycap pull force and laser-engraving parameter optimization.

• Pull-force prediction: the presentation explicitly lists Jason Hung as a participant. Material, opening dimensions and force data were explored to support design decisions and reduce trial and error.
• Cross-functional case: the key-pressing project reported an offline experimental metric increasing from 200 to 1,113 pcs per operation. This is a team-reported result for a specific operation, not a whole-line capacity increase.
• Implementation approach: connect shop-floor problems, data preparation, modeling, parameter recommendations and offline validation, distinguishing validated results from cost estimates.

Read the research and source notes

Jingtek E-bike App & IoT Service System

Product Planning & Entrepreneurship Research | E-bike / IoT

Draws on product development and App / administration architecture planning during the co-founder / executive vice president role.

Explored integration of e-bike hardware communications, mobile applications and administration systems, extending product capabilities into online transactions, after-sales services and user communities.

• Scope included vehicle anomaly detection, repair locations, online shopping, member and product administration, and customer-behavior and sales analysis.
• Planned community interaction, test rides and loyalty points to connect product services with market development.
• Explored GPS positioning, rental geofencing, communication compatibility and OTA update requirements.

Read the research and source notes

Agentic Workflows & Manufacturing Governance

Continuing Study | Agentic Workflow

Studied Morris Fan’s lecture materials to explore workflow integration and governance for private enterprise AI agents.

Used The Practice and Challenges of Agentic Workflows as a study reference, focusing on manufacturing problem definition, data and tool integration, human review and traceable decision processes.

• Studied task decomposition, knowledge retrieval, tool execution, result validation and human review.
• Examined ERP / MES / SCM integration, RAG, MCP, tool permissions, exception handling and audit trails.
• Explored equipment troubleshooting, scheduling, demand forecasting and supply-chain early warning as application scenarios.

Read the research and source notes

Expertizing Engineering Know-how & AI-Assisted Technical Review

Knowledge Assetization | 2026

Independently executed: know-how extraction, expert role definition, review process design, and human–AI division-of-labour validation

Converted 10+ years of embedded power and motor-control experience (commutation truth tables, protection criteria, CAN/J1939 protocol design) into a packageable, distributable professional expert role, so the know-how no longer leaves with the individual; validated against real firmware a division of labour where “AI runs the standardized first-pass review, humans own risk adjudication” — the first AI pass surfaced 3 safety-critical defects.

• Knowledge assetization: Extracted a decade of motor-control know-how (BMS protection thresholds, commutation truth tables, six-step to FOC migration paths, CAN and J1939 addressing practice) into a structured professional role covering four classes of executable rules — capability boundaries, workflow, output conventions and prohibited actions — deliverable to internal engineers or external advisors.
• Portability of judgment: Converted heuristics that previously lived only in personal experience and code comments (e.g., “protection thresholds must never be adjusted ad hoc,” “measured results and inferences must be labelled separately”) into explicit rules, reducing dependence on specific individuals for knowledge transfer.
• Human–AI division of labour, demonstrated: The AI independently surfaced 3 safety-critical and 1 high-risk defect in the first pass and built its own verification scripts covering bit-order comparison and field-vector derivation — reproducible, objective findings. Severity adjudication and solution trade-offs remained with the human, shifting senior effort from line-by-line checking to defining standards and final adjudication.
• Quality discipline and risk closure: A finding counted as settled only when reproduced by two independent methods, with verification scripts required to parse parameters directly from source rather than by manual transcription — which caught one transcription misjudgement. Findings graded by personnel and line safety into safety-critical / high-risk / general recommendation, each mapped to an assignable, verifiable corrective action, with items requiring hardware confirmation explicitly flagged together with the test method.

Read the research and source notes

E-Vehicle Controller Embedded Development

Ongoing Development | Embedded Systems

Personal Technical Research & Development (ongoing)

Building on 10+ years of embedded development experience, ongoing firmware and tooling development for an e-bike controller (three-phase BLDC motor control): motor control algorithms, migration of production-proven protection logic, communication protocol design, Windows tuning tools and AI-assisted diagnostics over logged operating data.

Read the research and source notes

E-Vehicle Battery Management System (BMS) Development

Ongoing Development | Battery Management

Personal Technical Research & Development (ongoing)

End-to-end design of a 36V (10S) lithium battery management system: analog front-end (AFE) selection, host MCU and CAN architecture, safety-gated regenerative braking with advance negotiation, and online terminal contact-resistance monitoring — with full schematic, pin-level wiring table and CAN protocol draft delivered.

Read the research and source notes

Industrial Edge Gateway (Hardware + Software + AI)

Ongoing Development | Industrial IoT

Personal Technical Research & Development (ongoing)

Born from a universal manufacturing pain point: legacy equipment (RS-232 proprietary-protocol meters, old PLCs) holds data that cannot be extracted, and vendor-based integration is costly with no guarantee of success. The direction of this project is a hardware-software-AI system that turns proprietary protocol integration from custom development into configuration — an in-house low-cost edge node board for physical access, a JSON protocol-description engine that parses any proprietary frame format with zero code changes, and AI that generates protocol profiles and alarm rules from packet samples and plain-language descriptions, connecting old machines to data platforms quickly and affordably.

Read the research and source notes

Smart Door Lock Product Definition & Selection (NPI)

Product Definition & Selection | Smart Lock

Requirement definition, solution selection and cost engineering — independently delivered (design phase, ongoing)

Targeting one hardware platform shared across three scenarios — home, apartment and small B&B — redefined the requirements for a full-featured smart door lock (WiFi / Bluetooth / key card / PIN / fingerprint / doorbell / App / smart-home ecosystem), and completed the ecosystem route decision, main-MCU and peripheral selection, the unit cost structure and mass-production cost engineering, plus the architecture plan for a hotel-grade complete system (lock + gateway + card encoder + management software); on top of existing Bluetooth lock assets, planned auto double-locking and tamper warning as differentiating features.

Read the research and source notes

2026 NTUT EMBA 12th Glocalynergy — Overall Champion

2026-04-18 – 2026-04-19

Served as captain of A2 Hantian Shenniu, leading the cross-disciplinary team to the overall championship among 15 teams. The event brought together more than 200 EMBA participants from eight classes, combining practical challenges, case analysis and strategic decision-making.

Economic Daily News coverage · April 23, 2026

2013 TAIPEI CYCLE d&i awards | Smart E-Box & APPs

2013

Product / team recognition for Smart E-Box & APPs, presented as part of the E-bike product development and entrepreneurship experience. The supplied award record credits Shanghai Yiten Electronics Co., Ltd. for design and Zhida C.E International Co., Ltd. as manufacturer. The award record and ceremony photograph are shown below.