We Build Tools That Actually Work

Started in 2019 when a small accounting firm in Taichung asked if we could automate their invoice processing. That project taught us something important: businesses don't need fancy AI buzzwords. They need scripts that run reliably at 3am and bots that don't break when someone changes a spreadsheet column.

How We Got Here

Our first client saved 14 hours per week with a Python script that parsed supplier emails. Nothing revolutionary, just solid code that handled their specific workflow. Word spread in their business network, and we suddenly had more projects than we could handle.

By 2021, we'd built automation systems for manufacturing shops, logistics companies, and retail chains across central Taiwan. Each project taught us new edge cases and failure modes. That hands-on experience matters more than any certification.

Today we're based in Nantou, working with businesses that range from family-owned operations to regional enterprises. We still write every line of code ourselves and test everything on real data before deployment.

Development workspace showing coding environment and testing setup

What Drives Our Work

Practical Solutions

We skip the theoretical frameworks and build tools that solve your actual problems. If it doesn't save you time or prevent errors, we won't suggest it.

Clear Communication

No technical jargon in our project updates. We explain what we're building, why it matters, and what could go wrong in plain language that makes sense.

Long-Term Support

Systems break when businesses evolve. We stick around to adjust scripts when your workflow changes or your vendor updates their API without warning.

Project Evolution

2020

Email Processing Systems

Built our first batch of email parsing bots for procurement teams. These systems monitored supplier communications, extracted pricing data, and flagged delivery changes. Still maintaining three of those original installations today.

2022

Inventory Automation

Developed warehouse monitoring scripts that synced stock levels across multiple sales channels. The tricky part wasn't the code but handling inconsistent product IDs between different platforms. Learned a lot about data validation that year.

2024

Custom Integration Tools

Started connecting legacy accounting systems with modern e-commerce platforms. These projects require understanding both old database structures and new API standards. It's messy work but incredibly valuable for established businesses.

2025

Expanding Capabilities

Now working with machine learning models for document classification and chatbots for customer service. We're careful not to oversell what these tools can do, but when applied correctly they handle repetitive tasks remarkably well.

Meet the Developer

Kęstutis Bakker, Lead Developer at SmartCortex AI

Kęstutis Bakker

Lead Developer

Started programming in university while studying industrial engineering. Spent five years at a manufacturing software company before going independent in 2019. Enjoys debugging edge cases and finding elegant solutions to messy data problems.

Outside work, I maintain several open-source Python libraries and occasionally speak at local tech meetups about automation best practices. Always happy to discuss project requirements over coffee.

Discuss Your Project