Skip to main content
JinzaiLab Ching Hsu's AI automation studio

Services

Start with the work you need to improve

I design around your current process, data, users, and constraints, then select technology that your team can maintain.

AI workflow automation and internal tools

For teams moving the same information between spreadsheets, email, cloud storage, and manual steps.

Examples: Data validation, email reporting, information collection, content delivery, and cloud backup.

Common problems

  • — Repeated copying and data cleanup
  • — Missed steps across multiple systems
  • — Processes known by only a few people
  • — Inconsistent output formats

Typical deliverables

  • — Workflow mapping and automation boundaries
  • — OpenAI, Gemini, Google, LINE, and other API integrations
  • — n8n workflows, Python batch tools, or small web tools
  • — Error handling, logs, deployment notes, and handover documentation

Document intelligence and research automation

For work involving large volumes of reading, classification, evaluation, comparison, and reporting.

Examples: Pitch deck evaluation, application document checks, CSV validation, research QA, and content strategy.

Common problems

  • — Information spread across file formats
  • — Inconsistent review standards
  • — Research that is hard to reuse
  • — Reports that take too long to prepare

Typical deliverables

  • — PDF, PPTX, DOCX, image OCR, and archive parsing
  • — Evaluation frameworks, rule checks, and LLM-assisted review
  • — Human review points and traceable output
  • — Markdown, Google Docs, Sheets, or HTML reports

AI-assisted media workflows

For teams handling large amounts of footage, subtitles, metadata, output formats, or media delivery.

Examples: Automated editing, stock-footage metadata, subtitle translation, batch video generation, and delivery.

Common problems

  • — Long chains of repetitive processing
  • — Manual caption and metadata work
  • — Batch generation without reliable monitoring
  • — Frequent delivery-format errors

Typical deliverables

  • — Speech recognition and subtitle pipelines
  • — Automated editing, transcoding, and resizing with FFmpeg
  • — Video description, classification, and keyword generation
  • — Vidu, Veo, Google Drive, FTPS, and other media API integrations

A project can start with one focused question

I can help with discovery, technical feasibility, one integration, an existing workflow, or a complete tool. You do not need a finished specification before contacting me.

Proof of concept

Test the model, API, cost, and output quality with real data before committing to a full build.

Workflow or tool development

Design the interface, data flow, error handling, deployment, and a usable first release.

Existing system improvements

Improve reliability, cost, performance, prompts, documentation, or human review points.

Not sure which approach fits your workflow?

Send me a short note about your current process and the outcome you need. I will help you identify a practical starting point.

Tell me about your project