Currently translating customer problems into working products

I materialize the vision before the solution is designed.

TODAY·CPO at dorje.ai

“An AI assistant agent for every business”

AI AGENTS

Customizable AI agents for enterprise and SMEs

Customizable AI agents across the market — from SME customer-service chatbots to enterprise agents with persistent memory, per-user isolation and end-to-end observability. Load-tested for exactly-once processing under real service quotas.

— Enterprise: AgentCore architecture — memory, isolation, observability— Proven zero message loss / zero duplication at ~0.6s steady-state— SME: full-stack chatbot SaaS — agents, cloud backend, web, Stripe
2026

“What if AI could read and file your invoice?”

AP AUTOMATION

AI-powered accounts payable

Customer-deployed accounts-payable automation: reads messy, multi-page invoices — attachments and all — and books them straight into Xero. Built the live POC, own design and delivery of the production build. Previously known as Iris.

— Cut per-invoice handling ~70% — ~10 min → ~3 min— Live with 2 customers— Built for ~1,000 invoices/month projected throughput

“I told our story on AWS stage”

GUEST SPEAKER

Sharing the learnings and industry insights on Intelligent Document Processing (IDP)

Selected as an AWS partner speaker to a 100+ audience in Bangkok — presented automating intelligent document processing with Bedrock, representing Dorje.ai to customers and the AWS field.

— AWS partner speaker, 100+ audience in Bangkok— Automating IDP with Amazon Bedrock— Represented Dorje.ai to customers and the AWS field
2025

The pivot — back to school, forward into business.

NUS BUSINESS SCHOOL · [ MSc IN MANAGEMENT ] · 2024

“The face of the programme for the region’s top students”

STUDENT REPRESENTATIVE

Representing NUS to Southeast Asia’s best

An NUS initiative bringing top-performing students from across Southeast Asia to tour campus over three days. I held two roles at once: activity lead for the open house, and one of four panel members chosen to share student life and study on campus.

— Activity lead for the 3-day open house— 1 of 4 selected to the student panel— Hosted top students from across Southeast Asia
2024

Proving driverless vehicles safe for Singapore’s roads.

ST ENGINEERING · [ SENIOR ENGINEER ] · 2019—2023

“18,000 labelled images an hour — from a digital twin of Singapore”

GEN2 6.8m BUS

Teaching a self-driving bus to see, in simulation

The Gen2 6.8m bus’s detection models needed huge amounts of labelled data — the bottleneck on the certification path. I built a high-fidelity simulation: a digital twin of Singapore’s One-North district, matched to the bus’s real sensor rig, that renders photoreal images and the matching in-sim point cloud. Because the sim knows where every object is, each frame comes out perfectly labelled — no manual tagging.

— Digital twin of One-North, matched to the real sensor config— Photoreal images + in-sim point cloud, auto-labelled— ~18,000 labelled data points/hour, supplementing real training data

“Certifying a 12m driverless bus — in simulation first”

AUTONOMOUS BUS

Singapore’s first 12m autonomous bus

Assisted on the 12m autonomous bus’s CETRAN certification — the M1 milestone, then the M2 Software-in-the-Loop stage, validating the stack in simulation ahead of the road. The bus went on to public trials in Jurong.

— CETRAN M1, then M2 (Software-in-the-Loop)— Simulation-first validation— Public trials in Jurong
2022

“Singapore’s first autonomous road sweeper — onto the road”

AUTONOMOUS ROAD SWEEPER

Singapore’s first autonomous road sweeper

Led the company’s internal test team preparing Singapore’s first prototype AV road sweeper for its CETRAN M1 milestone — the readiness work ahead of the authority’s certification, liaising with the Land Transport Authority (LTA) through the process. Automated the test process, cutting test-case prep from 2 weeks to 1 day.

— Led internal-team preparation for the M1 milestone— Liaised with LTA through certification— Test-case prep 2 weeks → 1 day (~10×)

“The bus I first tested as an intern — into public service”

AUTONOMOUS BUS

From intern’s project to public shuttle

A continuation of my internship: I stayed with the 6.8m autonomous bus into the full-time role, running its public-readiness trials. The bus went on to serve the public as an on-demand shuttle on Sentosa island.

— Continued from internship into the full-time AV role— Ran public-readiness trials— Became a public on-demand shuttle in Sentosa
2019

Where it all started — an engineer, oblivious to the adventure ahead.

CHULALONGKORN UNIVERSITY · [ BEng MECHANICAL / AUTOMOTIVE ] · 2015—2019

“Measuring how humans drive in a convoy — before the machines take over”

FINAL-YEAR PROJECT

Vehicle platooning research

Researched platooning — vehicles driving in a tight convoy — for low-speed private cars. Measured two drivers’ profiles in convoy: reaction time, time-to-collision (TTC) and speed profile, to characterise human following behaviour.

— Low-speed private-car platooning study— Measured reaction time, TTC and speed profile— Characterised human following behaviour

“An unexpected journey”

INTERNSHIP · SINGAPORE

The internship that became a career

Joined ST Engineering in Singapore as a testing-team assistant for the 6.8m autonomous bus in Sentosa — the placement that led straight to the full-time AV role, where I stayed with the same bus.

— Testing-team assistant, ST Engineering Singapore— 6.8m autonomous bus in Sentosa— Led directly to the full-time AV role

“Commissioned to model and 3D-print a hydro power station”

FREELANCE

Srinagarind Dam diorama

Commissioned by NICT to model and 3D-print a diorama of the Srinagarind Dam hydroelectric power station — its control box and the station buildings.

— Commissioned by NICT— 3D modelling and printing— Srinagarind Dam — control box and station buildings
2015
WHERE IT BEGAN