AI PRODUCT · WORKFLOW · QUALITY

Native Thai • Working English • Bangkok, Thailand • Remote

Taninnuth Warittarasith

AI ที่เชื่อถือได้ เริ่มจาก judgment ที่ชัดเจน

AI-Assisted Product Builder • AI Workflow & Quality Specialist

ผมเปลี่ยนปัญหาธุรกิจให้เป็นระบบ AI-assisted ที่ใช้งานได้จริง ด้วยการกำหนดกฎ exception เกณฑ์คุณภาพ และขอบเขต human review ให้ชัด ก่อนใช้ AI coding agent ช่วยเร่ง implementation.

ผลงานที่เลือก

ระบบที่ออกแบบจากปัญหางานจริง

สาม case study ที่แสดงการออกแบบ workflow, การใช้ AI เฉพาะจุด และ quality control ที่ตรวจสอบได้.

01

Internal

LINE OA Cafe Kiosk Ordering Engine

ระบบสั่งซื้อผ่าน LINE ที่ออกแบบด้วย state, failure handling และ multi-tenant workflow

WorkflowState MachineIdempotency

ปัญหา

เมื่อ order สลิป คิว และสถานะรับสินค้าอยู่กระจายในแชตและหลาย integration จะเกิดข้อมูลซ้ำ สถานะไม่ตรง และ notification ตกหล่นได้ง่าย.

สิ่งที่ผมทำ

ออกแบบ order state machine แบบ store-scoped พร้อม LIFF flow, Supabase live mode, admin RBAC, idempotency สำหรับ create/slip/webhook/action, failed-delivery replay และ tenant bootstrap.

02

Prototype

SlipLedger

Accounting workflow แบบ rule-first พร้อม optional AI และ human review

Human-in-the-loopRule-firstAccounting Workflow

ปัญหา

Statement หลักฐาน การจัดหมวด และการส่งต่องานบัญชีมักกระจัดกระจาย และการให้ AI ลงบัญชีอัตโนมัติโดยไม่มี traceability/review มีความเสี่ยง.

สิ่งที่ผมทำ

ออกแบบ rule-first suggestion + vendor memory + optional AI พร้อม transaction review การผูกหลักฐาน DR=CR validation ledger/trial balance/reconcile RBAC และ export pack.

03

Prototype

Lead Cap / Name Card Keeping

AI extraction นามบัตรลง Google Sheet ของผู้ใช้โดยออกแบบเรื่อง privacy ตั้งแต่ต้น

AI ExtractionPrivacyOCR

ปัญหา

การพิมพ์นามบัตรจากงานแฟร์ช้า แต่การรวมรูปและข้อมูล lead ทั้งหมดไว้ในระบบกลางก็เพิ่มความเสี่ยงด้าน privacy และ operations.

สิ่งที่ผมทำ

ออกแบบ LINE/LIFF identity, Google OAuth/Sheets, image gate ก่อน OCR เต็ม, duplicate check, quota/purchase state และ data model ที่ไม่ persist raw card image/lead record ในฐานกลาง.

ผลงานเพิ่มเติม

Combat Performance Fueling — Evidence & Rule Architecture

Traceability จากหลักฐานสู่กฎและ professional review ก่อน automation

Case Study

AI Intake Assistant

Structured intake และตรวจข้อมูลที่ขาดก่อนส่งต่องานให้ AI หรือคน

Prototype

Sales Proposal / Export Document Generator

Workflow สร้าง proposal และ PDF จากข้อมูลที่มีโครงสร้าง

Prototype

สิ่งที่ผมนำมาช่วยทีม

ใช้ judgment ที่ชัดเจนตั้งแต่ evaluation ถึง implementation

ผมทำงานในจุดที่ AI quality, operational workflow และ practical system ต้องเข้าใจง่ายและ verify ได้.

01

AI Evaluation & Quality

ผมประเมิน AI output เทียบกับเป้าหมาย ข้อจำกัด และบริบทที่ชัดเจน พร้อมบันทึก feedback ที่นำไปแก้ไขต่อได้.

  • เปรียบเทียบ Output
  • เกณฑ์คุณภาพ
  • ภาษาและบริบทไทย

02

Workflow & System Design

ผมเปลี่ยนงานธุรกิจให้เป็น state, rule, integration และ exception path โดยใช้ AI coding agent ช่วยเร่ง implementation.

  • Requirement และ Rule
  • Failure-aware Workflow
  • Practical Integration

03

Human Review & Verification

ผมให้คนยังรับผิดชอบการตัดสินใจที่ไม่แน่นอนหรือมีความเสี่ยง พร้อมเพิ่ม validation, traceability, test และจุด handoff.

  • Human-in-the-loop
  • Traceable Decision
  • Test และ Smoke Check

Background

เริ่มจากพื้นฐานวิศวกรรมด้าน Design และ Product ก่อนต่อยอดสู่การทำธุรกิจ และงานอิสระด้าน AI-assisted product และ quality.

2007–2009

Toshiba Carrier · Design Engineer

2009–2012

Seagate Technology · Product Engineer

2012–2013

Hana Microelectronics · Product Engineer

ปัจจุบัน

Business Owner · Independent AI Work

ติดต่อ

เปิดรับโอกาสงาน Remote ด้าน AI Quality และ Workflow

สำหรับงาน Generative AI Evaluation, Human-in-the-loop Operations หรือ AI-assisted Product Work สามารถติดต่อผมได้โดยตรง.