Tai VONG
Discuss a project
Work Projects Services Writing About Discuss a project

Available for selected consulting, contract & fractional work

Tai VONG

Backend & Platform Engineer · Technical Lead

I help product teams diagnose difficult backend failures, stabilize Go services, and turn fragile systems into observable, production-ready platforms.

Discuss a project View selected work
  • Go & gRPC
  • Distributed systems
  • Production reliability
  • Technical leadership

SYSTEM ONLINE 10.7769°N 106.7009°E Ho Chi Minh City

Selected work

Chosen for what became possible afterwards

Each of these started with somebody blocked — an operations team getting four different answers to the same question, a developer rebuilding the same service skeleton for the fourth time, a search team stuck on an Elasticsearch version they could not leave. Below: who they were, what was in the way, and what they could do once it was not.

Own project 2026 Active

grpcx

Who it’s for
Backend teams standing up new gRPC and REST services.
What was in the way
Every new service re-implemented the same plumbing — server lifecycle, structured errors, i18n, auth, observability — slightly differently, so how reliable a service was depended on who happened to write it.
What I did
Built and maintain an open-source Go framework that wires those concerns in once and exposes them the same way across gRPC and REST.
What became possible
A team can start a service with the reliability primitives already in place, instead of bolting them on after the first production incident.
View repository (opens in a new tab)

Personal side project 2025 Active

Yuzu Money

Who it’s for
People who want savings to earn yield without learning how the underlying protocols work.
What was in the way
Yield products are usually only legible to people who already understand the mechanics underneath them, which puts them out of reach for everyone else.
What I did
Backend engineering on this personal side project — the services behind accounts, balances, and yield tracking.
What became possible
Saving and earning are presented as ordinary money movement rather than as protocol operations the user has to reason about.
Visit yuzu.money (opens in a new tab)

Sanitized production case 2024 Complete

One business rule, four different answers

Who it’s for
An operations team whose daily work depended on generated documents agreeing with each other.
What was in the way
A configuration-dependent business rule was applied correctly in one path but silently dropped from other generated outputs, so the same request could produce different answers depending on where you looked.
What I did
Traced the rule through calculation, persistence, preview and export, named the invariant all four paths were supposed to share, and added regression coverage that compares outputs against each other.
What became possible
Every output settled on one verified rule, and the whole failure class became something tests catch rather than something customers report.

Sanitized production case — no public link

Own project 2021 Stable

go-bootstraper

Who it’s for
Developers starting a new Go service under time pressure.
What was in the way
The first day of every new service went to recreating the same skeleton — layout, configuration, Makefile, Docker — and the small differences between everyone's version compounded into real inconsistency later.
What I did
Built a bootstrapper that generates a standardized Go project in one step.
What became possible
New services start from a shared, reviewed structure, so the interesting work starts sooner and the boring parts stop drifting apart.
View repository (opens in a new tab)

Co-authored 2020 Published

High-performance multi-face tracking system

Who it’s for
Computer vision researchers and engineers building multi-face tracking on video.
What was in the way
Occlusion and missed detections fragment identities, so a single person's trajectory breaks into several unrelated tracks.
What I did
Co-authored a framework combining detection-tracking with tracklet-to-tracklet association, running semi-online so it trades a short delay for a decision it can stand behind.
What became possible
Published in IJACSA as a peer-reviewed system that runs at real-time speed with a two-second decision delay, giving other teams a reproducible approach to the fragmentation problem.
Read the paper (opens in a new tab)

Contribution to another project 2020 Complete

Vietnamese analysis for Elasticsearch

Who it’s for
Teams running Vietnamese-language search on Elasticsearch, and the maintainer of the analyzer plugin they depend on.
What was in the way
The community Vietnamese analyzer had fallen behind Elasticsearch releases, so upgrading the cluster meant losing Vietnamese tokenization.
What I did
Contributed a merged upgrade bringing the plugin to Elasticsearch 7.5.1, plus a separate follow-up proposal for lowercase tokenization that was closed without merge. The project belongs to its maintainer — these were patches to their work.
What became possible
Vietnamese-language search users could move to the then-current Elasticsearch 7.5.1 release without giving up the analyzer they depended on.
View merged version upgrade (opens in a new tab)

These 6 best show the pattern. The full archive holds all 56 safely presentable entries — developer tooling, research, patches to other people’s projects, and the coursework where the habits started.

View all projects

Ways to work together

From a five-day diagnostic to ongoing technical direction

Start here

Backend Reliability Diagnostic

Focused 5-day investigation

A focused five-day investigation into a backend system that's failing in ways nobody can quite explain. You get a failure map, prioritized findings, a concrete remediation plan, and a handoff your team can execute without me.

  • Failure map of the system
  • Prioritized findings, ranked by risk and effort
  • Concrete remediation plan
  • Handoff walkthrough with your engineers
Discuss a project

Go & gRPC Production Hardening

2–4 week sprint

Takes a Go or gRPC service from "it mostly works" to production-hardened — correctness, observability, failure handling, and operability.

  • Correctness & edge-case coverage
  • Observability wired in — metrics, tracing, logs
  • Failure handling & graceful degradation
  • Operability handoff docs & runbooks

Fractional Technical Lead

Part-time, ongoing

Part-time technical direction for teams that need senior judgment without a full-time hire.

  • Architecture & technical decisions
  • Delivery systems & engineering process
  • Mentoring & team growth

Also available for selected contract engagements, technical advisory, speaking & workshops, and expert calls — reach out with what you're working on.

How I work

The same loop, every time

  1. 01

    Observe

    Read the signal before touching code — logs, traces, metrics, and the shape of the failure.

  2. 02

    Reproduce

    Recreate the failure on demand. If it can't be reproduced, it can't be trusted as fixed.

  3. 03

    Find the invariant

    Identify the assumption the system is silently violating.

  4. 04

    Fix the system

    Fix the invariant, not just the symptom, so the failure class doesn't come back.

  5. 05

    Verify in production

    Confirm the fix holds under real traffic, not just in a test environment.

Writing & open source

Notes on Go, gRPC, and distributed systems

Medium

Implementing common Go HTTP middleware

Middleware as a wrapper that runs before and after the main handler, keeping business logic free of cross-cutting concerns.

Read the article

Medium

Using buf.build to generate your gRPC codes

A tour through migrating an old protobuf/gRPC codegen pipeline to buf, based on hands-on research and practice.

Read the article

Research paper

A high-performance multi-face tracking system

Detection-tracking and tracklet association in a semi-online framework designed to run at real-time speed.

Read the paper
More on GitHub

About

Tai VONG

Technical Manager at Gearment, leading engineering across product and platform domains — architecture, delivery, and the teams that ship both.

My background is hands-on: backend and platform engineering, Go and gRPC, distributed systems, and the operational work of keeping production systems observable and reliable.

Earlier research background in AI and computer vision occasionally shows up in how I approach problems, though it isn't the focus of my work today.

Ho Chi Minh City, Vietnam

Have a backend problem worth a second opinion?

Tell me what's failing and what you've already tried. I'll reply with whether I think I can help.

chitai.vct@gmail.com Connect on LinkedIn