about · Taipei, Taiwan

Applied math, turned into software you can measure.

I started out pricing risk as an actuarial student in Thailand, spent three years wrangling data for banks and telcos in Bangkok, moved to Taiwan for a master’s in reinforcement learning, and now build decision-making systems: RL agents, world models, and LLM workflows. The through-line is the same question every time: how would we know if this actually worked?

Open to ML / AI Engineer and Forward-Deployed Engineer roles. Taipei, UTC+8. Email sakkarink12@gmail.com.

Résumé (PDF) ↓

§ 01

The path so far

01 · 2017
Actuarial science
Mahidol University · TH
Learned to price risk, and to distrust a single number.
02 · 2017 – 2020
Data engineering
Bangkok · TH
Hadoop for a Canadian bank, an AWS migration, models for Isuzu. Mostly on-site with clients.
03 · 2020 – 2024
MSc, RL research
NTHU · TW
Moved to Taiwan. Inverse RL for autonomous driving.
04 · since 2023
AI tooling + fund systems
SystemWeb · TW
Claude Code plugins and agent memory for the team; C# systems for fund administrators.
05 · now
Research in public
nights & weekends
World models, RL, agentic LLM systems, co-built with Claude.

§ 02

Experience

Most recent first. The current role gets the space; the Bangkok years get the highlights.

Oct 2023 — now · Taipei, Taiwan

SystemWeb Technologies

Software Developer

active

I build the AI tooling our engineering team works with every day, on top of the C# fund-administration systems we ship.

Claude Code plugins published internally: 3
per-feature development time: days → hours
engineers using them daily: 3–4
documents in the agent memory graph: 500

Architected and published 3 plugins to our internal Claude Code plugin marketplace, for task execution, ideation, and AI-native SDLC onboarding for engineers new to AI-assisted work. Per-feature development went from days to hours, and 3–4 engineers use them every day.
Designed a graph-based persistent memory layer for our agent system, replacing ad-hoc context handling. It indexes 500 documents so agents can recall what happened in earlier sessions.
Underneath it all: the C# Transfer Agent System for asset- and wealth-management clients (and the Portfolio Management System before it), built with a 5–10 person team split between Thailand and Taiwan.

Claude Code · Agent memory (graph) · LLM multi-agent systems · C# · JavaScript · SQL

2017 — 2020 · Bangkok

Data engineering, mostly on-site with clients

Before AI engineering I spent three years embedded with other people’s data: a Canadian bank’s Hadoop cluster, Thailand’s largest mobile operator, a real-estate developer’s move to AWS, Isuzu’s campaigns. That’s where I learned the forward-deployed basics: listen first, and assume the data is never quite where the slide deck says it is.

Feb — Jun 2020
Data Analyst · Tri Petch IT Solutionsfor Isuzu Motors
Predictive models for Isuzu’s marketing campaigns, plus the databases and cloud services behind them.
Python · SQL · Cloud
Oct — Dec 2019
Data Engineer (contract) · Pruksa Real Estate
Designed the data workflows and migrated on-premises data to AWS for the DataOps team.
AWS · ETL
Jul — Sep 2019
Data Analyst (contract) · Advanced Info Service (AIS)
Regional data-business requirements for Thailand’s largest mobile operator (40M+ customers).
Analytics
Mar 2018 — May 2019
Hadoop Developer · ATA IT Limitedfor National Bank of Canada
ETL pipelines and semi-structured data integration on Hadoop for National Bank of Canada.
Hadoop · Spark · ETL
Jan — Feb 2018
Junior Data Scientist · GameSpace
Data analysis and databases for a retail client with 10+ branches.
Sep — Nov 2017
Data Analyst · Betimes Solution
ETL, interactive dashboards, and prediction models for government-sector clients.
ETL · Tableau

§ 03

Education

2020 – 2024

MSc, Information Systems & Applications

National Tsing Hua University · Hsinchu, Taiwan

Thesis: enhancing autonomous driving with double critic + MCTS for inverse RL. Coursework in reinforcement learning, neural networks, and cryptography.

Feb 2017

BSc, Actuarial Science

Mahidol University · Bangkok, Thailand

Mathematics, statistics, finance, and economics: the applied-math base everything else sits on.

§ 04

Skills

Only what my work history or public repos can back up.

LLM & agentic systems

Claude Code plugins
LLM multi-agent systems
Agent memory (graph-based)
Prompt & context engineering
LLM evals
Hybrid frontier/local orchestration

Reinforcement learning

SAC / MaxEnt RL
PPO
World models (RSSM / Dreamer-style)
Multi-agent RL
Inverse RL

ML stack

PyTorch
TensorFlow.js
Gymnasium
MLX
Apple FoundationModels
CUDA (learning)

Data engineering

Apache Spark
Hadoop
ETL pipelines
AWS migration
SQL / PostgreSQL
Tableau

Software

Python
TypeScript / JavaScript
C#
C / C++
R
FastAPI
Next.js / React
SwiftUI

Infra & practice

Docker
AWS
Google Cloud
GitHub Actions
ADRs & pre-registration
Seeded, config-snapshotted runs

Languages

ภาษาไทย · native
English · professional · TOEIC 930
中文 · basic, and improving one menu at a time

§ 05

How I work

Plan first, read widely, then go explore

Most of my projects start life as a plan doc and a reading list. janus-chrysalis keeps notes on every paper it leans on. Then I try the new approach anyway, because that’s the fun part, and the plan tells me whether it actually made things better or just made them different.

Small steps, real numbers

Ship something small, measure it, then decide. I’d rather show you a modest number I can defend than a big one I can’t. That’s why every number on this site comes with its n, and why one of my favourite results is a null result.

Start from the person, not the tech

Figure out what they actually need first. So far that’s been fund administrators, a Canadian bank’s data team, Thailand’s largest mobile operator, and, for Fortuna, me squinting at my own credit-card statement. The model comes second.

§ 06

Off the keyboard

Brazilian jiu-jitsu

The most honest feedback loop I know: bad policy, immediate negative reward.

Cooking

Recipe testing is just hyperparameter search with better snacks.

Films

How I switch my brain off after a long training run. Somebody else’s story, somebody else’s decisions.

Want the short version? Say hi, in English, ภาษาไทย, or 中文 if you speak slowly.