Designing production systems @ TikTok

Marcus Peh.

Software Engineer @ TikTok · Backend & Frontend.

Scaling query infrastructure and data products at TikTok, with a focus on high-throughput systems and the interfaces that make them useful. Previously Google, ByteDance, and GovTech.

Resume GitHub LinkedIn

01 — Experience

Now, and before.

Currently focused on high-scale query infrastructure and data products at TikTok. Previously, engineering internships at Google, ByteDance, and GovTech.

Now · Current role

Software Engineer · Backend & Frontend

@ TikTok

Apr 2024 — Present

Singapore

Building high-scale query infrastructure and data products at TikTok. I improved the realtime user segmentation system to sustain over 2M peak QPS at approximately 30ms P99 with over 99.99% SLA, re-architected Hive-to-ClickHouse loading to reduce data readiness time by 73%, and designed a multi-region on-demand pipeline for cohort-level Creator Analysis with 30%+ user penetration.

Sustained over 2M peak QPS at approximately 30ms P99 with over 99.99% SLA by optimizing realtime user segmentation system cache access patterns.
Reduced data readiness time by 73% (5.5 hours → 1.5 hours) by re-architecting Hive-to-ClickHouse data loading.
Designed and built a multi-region, on-demand data pipeline for cohort-level Creator Analysis, reaching 30%+ user penetration.
Expanded Data Suite query capabilities with a visualization engine supporting direct ClickHouse queries and configurable virtual columns.

GoPythonTypeScriptSQLRedisClickHouseMySQLReact

Before · Engineering internships

Google

Software Engineering Intern · Google Pay

May 2023 — Aug 2023

Refined the onboarding verification flow on Google Pay’s merchant application (Android).

JavaProtocol BuffersgRPC

ByteDance

Backend Software Engineering Intern · Global Payments

Aug 2023 — Dec 2023

Improved channel rating API performance by 90%+ by precomputing data and parallelizing category calculations with Go goroutines.

GoMongoDB

GovTech

Software Engineering Intern · CrowdTaskSG

May 2022 — Jul 2022, Dec 2022 — Jan 2023

Reduced server query response time under load by 200× by restructuring API request handling and reducing redundant database queries.

TypeScriptReactJSPostgreSQL

02 — Side projects

Projects on the side.

Things I've built or actively maintain outside of work

01 / 03

Trading Algorithm Framework

Modular Go framework for running the same strategy logic across backtesting and live execution.

Reusable components cover market data, order execution, portfolio management, risk controls, scheduling, and historical backtesting.

GoBacktestingLive executionRisk controls

View on GitHub

02 / 03

Market Data Service

Provider-agnostic Python service exposing historical OHLCV and ETF constituent data across 15+ years through REST APIs.

Parquet-based caching and scheduled constituent snapshots reduce repeated upstream requests and keep historical backtests reproducible.

PythonFastAPIREST APIsParquet

View on GitHub

03 / 03

Spending Tracker (Telegram Bot)

Telegram bot that polls an email inbox for transaction records and auto-logs each one into a personal expense ledger.

A lightweight expense capture flow: transactions are pulled straight from email, so logging happens without another app.

PythonIMAP / Gmail APITelegram Bot APISQL

View on GitHub

03— Currently

Where my attention is.

Research and infra work happening alongside production systems.

Trading algorithm framework

A modular Go framework for backtesting, live execution, and risk controls.

A modular Go framework for backtesting, live execution, and risk controlsIn progress

Market data service

Provider-agnostic historical OHLCV and ETF constituent data over REST APIs.

Provider-agnostic historical OHLCV and ETF constituent data over REST APIsExploring

Spending tracker (Telegram bot)

Polls an email inbox for transaction records and auto-logs them to a personal expense ledger.

Polls an email inbox for transaction records and auto-logs them to a personal expense ledgerBuilding

MSc Computer Science

Studying at Georgia Tech while continuing to build at TikTok.

Studying at Georgia Tech while continuing to build at TikTokStudying

04 — About

In short.

I'm a software engineer at TikTok building query infrastructure and data products that perform under pressure.

I enjoy working on high-scale systems problems, optimizing critical paths, and turning data into products that teams can use to make better decisions.

Based in: Singapore
Focus: Query infrastructure & data products
Education: MSc CS · Georgia Tech (ongoing)
Undergrad: NUS · Highest Distinction

05 — Get in touch

Open to senior software engineering roles and systems conversations.

Especially interested in distributed infrastructure, decisioning platforms, and product-focused engineering teams.