Personal workflows
Reduce repetitive work, organize knowledge entry points and make AI a reliable everyday collaborator.
PROMPT → WORKFLOW→
Applied AI × Data Systems
Turn AI and data
into systems that
move work forward.
From problem framing, data and models to knowledge workflows and organizational adoption. I help research teams and companies build AI systems that can be executed, verified and continuously improved.
Dr. Jui-Ming ChangAI Transformation & Data Systems Consultant
Discuss a project ↗
00
Do not start from a tool list. Clarify the boundaries and target first, then decide which data, models and workflows are actually needed.
01 / TRANSFORMATION
AI adoption is not a
AI adoption is not a
one-time tool purchase.
It is a maturity path shaped by people, data, processes and risk boundaries—starting with small, verifiable steps.
Labs and teams
Create shared data structures, knowledge bases and collaborative workflows so information does not live in one person's head.
TOOLS → SYSTEM→
Organizations
Move from needs assessment and pilot validation to governance boundaries and a maintainable adoption path.
PILOT → ADOPTION↗
02 / SERVICES
Clear capability modules.
Clear capability modules.
Tailored to the problem.
Every engagement begins by diagnosing the need, current data conditions and success criteria—then selecting the right capability mix.
AI transformation &
workflow design
Assess the current state, identify high-value entry points and design an adoption path for labs, teams and organizations.
- Current-state and needs diagnosis
- Adoption roadmap
- Verifiable workflow prototype
Workflow transformation↗
Data &
ML systems
Build complete systems—from data structures to deployed models—that can handle real-world time series, noise and imbalanced data.
- Databases and data pipelines
- Feature engineering and model training
- Evaluation, alerts and iteration
Data to deployment↗
LLM
customization
Connect language models to organizational knowledge and tasks while setting clear boundaries for automation, verification and human judgment.
- RAG and knowledge retrieval
- Agentic workflow
- Fine-tuning and local model evaluation
Knowledge to action↗
03 / EVIDENCE
Results are not decoration.
Results are not decoration.
They pressure-test the method.
From real-world data and workflows to knowledge translation, the work demonstrates delivery at different scales.
ML models &
field applications
Start from the data and its physical meaning, then build models for time series, noise and imbalanced data that can be evaluated, deployed and iterated.
Rockfall / debris-flow monitoring
Anomaly detection and alerts
LINE rock classifier
Signal → decision
AI workflow
transformation
Reorganize scattered tools, manual curation and individual experience into traceable, repeatable knowledge workflows that can keep improving.
Research and knowledge work
Team collaboration
Human-in-the-loop
Transformation path
BoundariesData & knowledgeModel validation
AI education &
organizational adoption
Translate fast-changing AI concepts into shared language and methods that different audiences can understand, question and act on.
Talks and workshops
Research / education / government
Tools and workflow training
Knowledge → action
30+Talks &
workshops
workshops
03Sectors
Dr. Jui-Ming Chang
AI Transformation & Data Systems Consultant
I approach problems with a researcher's rigor, turn methods into working systems through engineering practice, and translate complexity into a next step teams can act on.
My work spans real-world sensing data, ML systems, agentic AI, knowledge workflows and cross-domain communication.
LinkedInBring the problem.
Let's find the next step.
