My Journey as an AI Engineer

Helped teach students at LEARN SATIT PATTANA SCHOOL how to build their own games on Roblox, along with other AI-powered game experiments.

Focused on getting kids shipping something end-to-end — from first idea to playing each other's games in class — and catching the reaction the first time their own character actually moved.

Reinforced that "teach kids to code" only works when paired with "make it fun first" — starting from games and AI lowers the fear-of-code barrier instantly.

Rewarding to share tech with the next generation and see how AI is already reshaping the way kids approach building anything digital.

Launched SPU Care — an issue reporting and tracking system for Sripatum University (February 13, 2026).

Led the design and development with the SPU AI Club team. Supports QR-code reporting with no sign-up, plus instant photo and location attachment.

Integrated AI to analyze and categorize reports, elevating the university's case management efficiency.

Users receive an automated tracking number and can monitor status in real time.

Invited to speak at One Stop Open House 2026 — a major education expo under the theme "Open Mind, Open Future", helping students explore their future paths.

My talk covered how AI is reshaping teaching, learning, and student development — and what the future of learning looks like when AI sits beside every learner as a co-pilot.

The event featured 70+ booths from leading universities across 3 zones, over 10,700 scholarships, and 3 full days of sessions — opened by the Minister of Higher Education, Science, Research and Innovation.

Watching students actively search for their path reminded me AI isn't this generation's competitor — it's the lever that will take them further than they currently imagine.

Won 2nd Runner-Up at the Amazon Kiro — Build AI-Powered Education Applications Hackathon 2026, hosted by SPU × AWS × Go Digit.

Our team built StepUp+, an AI tutor avatar for education, using Kiro, AWS Lambda, DynamoDB, and Amazon Bedrock for generative AI.

A 2-day format — full-day AWS workshops on Serverless, Containers, GenAI, and Agentic AI, followed by a build sprint and a demo in front of the judging panel.

Got hands-on with the full Agentic AI stack — Strands Agent, AgentCore, Knowledge Bases, Guardrails — while shipping something real with a team under real time pressure.

Won 1st Runner-Up at HAMMER CASE 2026, the business case competition by Hammer Club themed "AI for Help SMEs", representing SPU AI Club in a five-person team from Sripatum University's Faculty of Entrepreneurship and Faculty of Accountancy.

The problem: SMEs selling in Modern Trade can't see data or shopper behaviour at the shelf, so their marketing, product and business decisions are guesswork. Our answer was HEN (เห็น) – Smart Shelf Intelligence: a smart screen with a camera on each shelf, powered by an NVIDIA Jetson Orin Nano at the edge, that turns an ordinary shelf into full shopper intelligence and in-store ad space, as part of a full marketing loop for SMEs: AI-made ads → HEN shelf screens → the MEETA in-store promotion robot that shows promotions and invites shoppers to try products.

I owned the system architecture and the edge AI pipeline: YOLO detects people, ByteTrack follows their movement, InsightFace estimates age range and gender anonymously, and every image is deleted right after processing — privacy by design for PDPA. SPU AI Club backed the team with space, equipment and advisors.

Won 1st Runner-Up at LearnHack 2026 | AI Hackathon for the Next Generation (15–16 Aug 2026), co-hosted by Adam's House and the Faculty of Entrepreneurship, Sripatum University, with Lovable as a partner.

Together with high school students from La Salle School, our team built LearnSpace AI in a single day on Lovable: a learning platform that turns a teacher's own documents into lessons, quizzes and flashcards, gives students an AI assistant that answers from class material with citations, and shows teachers which students need help first.

We pitched it to a panel of three judges and won 600 Lovable credits. The real takeaway was running the full loop from a real classroom problem to a working prototype to a pitch, under hackathon time pressure.