Founder

From discovery to operations,
the founder is in every step

Mike Feng, founder and CEO of Sainso Technology

Mike Feng 馮志聖

Founder & CEO

Company
Sainso Technology Co., Ltd.
Based in
Taichung, Taiwan
Tax ID
62171540

Mike Feng spent ten years building full-stack, cloud, IoT and AI systems for Japanese and Taiwanese companies before founding Sainso Technology in 2026, an AI systems-integration company in Taichung, Taiwan. He works on every project himself — requirements, architecture, full-stack development, AI integration, cloud deployment and ongoing operations — so clients talk to the person who actually builds the system, not a sales intermediary. At all three of his Japanese employers he started as the only engineer and built the system from zero; at Trim and Terra Motors the development teams grew from him alone to 5–10 and 20–30 people.

Sainso delivery evidence

Delivery records from Sainso Technology. Figures come from the homepage case studies and carry their measurement dates; client names are kept confidential under NDA.

Tire retail chain · ERP + LINE store lookup

Connected the client's existing ERP so members get tier-based quotes over LINE; rolled out in 12 phases from 2026-06-16, with 159 existing member accounts approved (2026-09-26).

Service-business platform · LINE AI customer service

Each shop uses its own LINE Official Account; complaints, refunds and personal data always go to a human. A production demo store replied in 2 seconds on a real device (2026-09-14).

Sainso website · AI content editor

AI picks topics and writes bilingual articles that must pass a quality gate before publishing; 123 articles published in three months since 2026-06-28.

Sainso operations · on-call agent

24/7 scheduled checks produce daily operations and security reports; 4,178 reports since 2026-06-06 (counted 2026-09-30). Deployments and outbound messages always need human approval.

Career

Career & education

2026 – now

Sainso Technology · Founder & CEO

AI systems integration, LINE / ERP integration, cloud deployment and operations.

2025 – now

Hoteck (Taiwan) · AI Project Manager

Industrial HVLS fan maker: leads the onaiir airflow-simulation platform, wind-lab measurement design, production-line AI for motor-noise detection, and the test-system database.

2023 – now

Terra Motors (Japan) · R&D Technical Lead

Built Terra Charge, a cloud-based EV charging platform, from zero as its only engineer: charger control (OCPP, MQTT, AWS IoT Core), reservations, admin console, mobile apps and payments. The development team has grown from him alone to 20–30 people across Japan and India.

2019 – 2022

FLECT (Japan) · R&D Engineer

Sole developer on every project, with a Japanese colleague handling the client side: delivered AI, IoT and cloud projects for manufacturing, retail, insurance and entertainment clients, and researched AR, LiDAR and indoor positioning.

2016 – 2019

Trim (Japan) · CTO

Built mamaro, an IoT smart nursing room, from zero as a one-person team — sensors and circuit boards through to the cloud, plus sourcing and importing the units from manufacturers in China — and the MakeLocal community network for parents. The development team grew from him alone to 5–10 people.

Education

M.S., Computer Science and Information Engineering, National Ilan University (2014–2016)

Research in image processing. B.S., Computer Science and Information Engineering, Asia University (2010–2014).

Other roles

Work the founder did in other roles — not Sainso client results.

Production-line AI

Found the 50.28 Hz modulation shared by every noisy DC motor; one interpretable metric classified all 36 recordings from 6 motors correctly and tells impact-type from modulation-type faults to guide teardown.

Airflow simulation & measurement

Built the onaiir 3D airflow-simulation platform (fans, HVAC equipment and temperature fields computed in real time; read-only client links) and designed a fixed 16-probe anemometer array that measures 45 fan configurations in about 3 working days instead of 229 hours of point-by-point scanning.

Data platform at scale

Built distributed IoT ingestion on AWS for a construction-machinery maker and load-tested Amazon Aurora PostgreSQL with 3.5 billion records (about 1 TB).

100,000 concurrent users

Load-tested a real-time voice/video fan service for an entertainment company with 100,000 concurrent users on Amazon EKS.

Computer vision

Barcode AI that reads damaged, blurred, curved and glare-affected labels (Mask R-CNN); tire tread-wear diagnosis from photos; drone-based roof damage assessment.

Writing & conferences

16 technical articles on FLECT's engineering blog (2019–2021) on AR, LiDAR, indoor positioning and machine learning, for example:

International conferences (2015–2016): FC 2016 (IET/IEEE International Conference on Frontier Computing); International Conference on Science, Engineering and Technology Innovations, Osaka; Asian Conference on Engineering and Natural Sciences, Tokyo.

Expertise

Expertise

AI application architecture

RAG knowledge bases, multi-model routing, tool calling, evaluation flows, and data security design.

Enterprise integration

LINE, ERP, CRM, payments, membership systems, and internal APIs connected into maintainable data flows.

Cloud and platform engineering

AWS / GCP, CI/CD, monitoring, permissions, cost optimization, and reliability design.

IoT data applications

Device connectivity, edge collection, time-series data platforms, anomaly detection, and dashboards.

How we work

How we work

  • Fixed quotes by phase and scope — not billed per person-month.
  • Client data is used only for that project and never to train models.
  • Security processes designed against ISO 27001 controls; irreversible actions such as deployments and sending contracts always need human approval.
Chat on LINE