Sian Siong
(Edmond) Ng

Backend Developer · System Architect

I take systems from requirement to production.

Most recently: exactly-once reward delivery to 300K recipients per campaign, self-healing after pod failures, and a message-broker migration shipped with zero downtime.

Open to work — Back End Developer roles
Get in touch Download Resume

Signature Work

Case 01 — Exactly-once at scale

Exactly-once reward dispatch across a multi-pod cluster

Situation

A customer-engagement platform dispatches rewards for campaigns reaching up to 300K recipients each — roughly 1.5M dispatches at peak across 5 concurrent campaigns, processed by multiple pods in parallel.

Constraint

Message reordering and pod races could double-pay or drop rewards. Real money; retries are not free; a pod can die mid-dispatch at any moment.

Approach

Designed an epoch-gate + compare-and-swap (CAS) idempotent dispatch, backed by a 4-pass mark-first reconciliation orchestrator so interrupted campaigns self-heal without duplicate or lost sends.

Outcome

Exactly-once delivery held in production at full campaign scale — including through mid-dispatch pod deaths.

300K recipients / campaign · ~1.5M dispatches at peak · 0 duplicates

Case 02 — Zero-downtime migration

Pulsar → Temporal migration with zero downtime

Situation

The delivery engine ran entirely on Apache Pulsar consumers. Orchestration logic had grown hard to reason about and recover.

Constraint

The engine dispatches live campaigns daily — no maintenance window, no dropped or duplicated messages allowed during cutover.

Approach

Led a phased migration to a hybrid model: Temporal owns orchestration and recovery semantics, Pulsar consumers keep the high-throughput fan-out they are best at.

Outcome

Cutover completed with zero downtime and no delivery incidents; recovery paths became explicit workflow code instead of implicit consumer state.

0 downtime · 0 delivery incidents during cutover

Case 03 — 4.5 years of ownership

High-frequency trading infrastructure, end to end

Situation

At MyITS.co, thousands of automated trading bots executed high-frequency strategies across Binance, OKX and Bitget.

Constraint

Low-latency execution, real-money risk, and 24/7 markets — the system could never quietly be wrong.

Approach

Owned the backend for 4.5 years: spot/futures engines, a unified exchange gateway, indicator and backtesting frameworks, risk monitoring (anomaly detection, fund movements), and Prometheus/Grafana observability.

Outcome

A platform where strategies, risk controls and monitoring evolved together — built, operated and led by the same team that answered for it in production.

Thousands of concurrent bot executions · 3 exchanges · 24/7

How I Work

Design for failure first.

Every dispatch path assumes a pod can die mid-flight — that is why reconciliation is a first-class component, not an afterthought.

Prove it with numbers.

Correctness claims are verified at production scale: 1.5M budget deductions without over-spend, not a unit test and a hope.

Migrate without drama.

The Pulsar → Temporal cutover shipped in phases with zero downtime — big changes should be boring to operate.

Own the whole path.

From client requirements to architecture, delivery and mentoring — I stay accountable for what runs in production.

Demos

Experience

Nov 2025 – Present · Kuala Lumpur

Senior Backend Developer

SNSoft Sdn Bhd

Own end-to-end delivery of distributed backend services for a client-facing customer-engagement platform (multi-channel campaigns, rewards, budgeting) — see Case 01 and Case 02 above. Also established an L0–L3 layered architecture standard enforced by automated dependency-cruiser checks.

May 2021 – Oct 2025 · 4 yrs 6 mos · Malaysia

Backend Developer, System Architect & Team Lead

MyITS.co

Built and led the trading-bot backend covered in Case 03, plus the customer portal (Next.js web, React Native mobile) and the platform's public multilingual site.

Jul 2019 – May 2021 · Malaysia

Full Stack Developer

CCAR

Led development of an automotive marketplace platform (Carlist.my-style): listings, search, auth, payments and admin dashboards, end to end.

Skills

Core — I design and answer for these in production

NestJS / Node.js · TypeScript · MongoDB · Redis · Distributed systems & event-driven architecture · Idempotent / exactly-once processing · Temporal · Apache Pulsar · BullMQ

Working — I ship with these regularly

gRPC · WebSocket / Redis Pub-Sub · Prometheus & Grafana · Jest · dependency-cruiser · Next.js · Tailwind CSS

Familiar — I can navigate and contribute

React Native · Ant Design · Cloud infrastructure security · Business analysis

Get in touch

Open to backend and system architecture roles. Reach out by email or connect on LinkedIn.

[email protected] LinkedIn ↗ Resume ↓