DJ//DIGITAL_JEFFREY
IDENTITY/PERSONAL AI SYSTEM // DJ-01

Active personal project

Wen-Chia Yang

Building Digital Jeffrey

Jeffrey, Amplified.

I started Digital Jeffrey with a simple idea: if AI can reduce repetitive work and information overhead, I can spend more time solving harder problems, helping people around me, learning, and being with family. I am not building a digital replacement for myself. I am building systems that help me use my time and attention more deliberately.

Explore Digital ForgeMeet Digital JasperExperience behind the systemsOpen contact channel

MISSION / PERSONAL AI SYSTEM

Digital Jeffrey

AI should give time and capability back to people—not replace them.

Digital Jeffrey is my long-term project for bringing AI into everyday life without handing away judgment or responsibility. The goal is not to stay busier. It is to spend less time repeating work and reconstructing context, then use the time that comes back for deeper work, helping others, family, and learning.

I began with software engineering because it is where I can test the idea against real systems and observable results. That became Digital Forge. Parenting is the next domain: Digital Jasper helps two first-time parents organize what they are learning and prepare better questions while keeping parental and medical judgment human.

FORGE / ai-teamsOPERATIONAL

Digital Forge

A work system powered by ai-teams for requirements, planning, bug fixes, feature development, verification, incidents, releases, maintenance, backups, health checks, and technical-debt work.

CONTROLAI performs defined work. Humans retain review, production authority, and the veto.

Open the Digital Forge dossier →
JASPER / Parenting learning toolPROTOTYPE 1.0 / IN PROGRESS

Digital Jasper

A private learning tool that brings caregiving and growth records, study notes, selected public references, and physician-conversation notes into one context for organizing observations and preparing better questions.

CONTROLAI supports learning. Parents and professionals decide.

PRIVACY BOUNDARYThe website describes only the purpose and governance of Digital Jasper. It publishes no identifiable child data, raw records, medical details, physician-conversation content, precise timelines, photos, prompts, or provider logs. Any public example must use synthetic data.

0139Go servicesArchitecture responsibility
0214K+Registered playersJuly 2026 snapshot
031,000Concurrent playersValidated in one tournament
04370+GKE podsJuly 2026 snapshot
LIVE PRODUCT / ACTIVE DEPLOYMENTONLINE

FIELD SYSTEM // CP-01

CyberPoker

A live Texas Hold’em and social poker platform with 14K+ registered players as of July 2026, powered by a real-time, event-driven backend.

My role
Primary backend architecture and technical direction since Feb 2023
Core systems
Go · NATS · PostgreSQL · Redis · GCP · GKE
Open live productcyberpoker.online ↗

PROJECTS

/

FIELD OPERATIONS // 02

OPS-01VALIDATED

Real-Time Tournament Scale

150 → 1,000players across 100+ active tables

Mission / Challenge

A single tournament became noticeably slow at approximately 150 simultaneously connected players.

Action / Outcome

Combined database read/write separation, concurrent processing, Redis caching, higher NATS storage IOPS on GKE, and sustained WebSocket load testing to raise validated capacity to 1,000 players across 100+ tables; verified tournament completion and reward distribution.

SystemsGoWebSocketNATSPostgreSQLRedisGKE
OPS-02DELIVERED

Reliable Payment Integrations

3third-party payment providers

Mission / Challenge

Virtual-goods purchases required reliable confirmation across three providers despite duplicate and out-of-order callbacks.

Action / Outcome

Integrated all three providers with signature verification, Redis-backed idempotency, transaction-state controls, callback-order handling, and reconciliation support.

SystemsGoRESTRedisPostgreSQLIdempotency
OPS-03IN ACTIVE USE

Production Recovery & Prechecks

~1 HOURrecovery with no permanent data loss

Mission / Challenge

An unexpected GKE node-pool upgrade restarted pods before active tournament state had been backed up.

Action / Outcome

Diagnosed the restart path, restored tournament state within approximately one hour without permanent data loss, and introduced production read-only maintenance prechecks; destructive operations remain manual.

SystemsGCPGKEKubernetesBashPythonIncident Response
OPS-04GOVERNED

Human-Governed AI Delivery

30 MIN → <5 MINtypical API-document workflow

Mission / Challenge

Repetitive documentation and cross-functional AI use needed a consistent workflow with explicit human accountability.

Action / Outcome

Connected requirement clarification, architecture, planning, implementation, independent verification, and approval gates; reduced typical API-document work from approximately 30 minutes to under five and established fixed workflows across product management and marketing.

SystemsClaude CodeCodexJSON SchemaMarkdownHuman Approval
OPS-05DEV INTEGRATION TESTING

AI-Governed Feature Delivery

2 FEATURESdevelopment complete and deployed to dev integration testing

Mission / Challenge

Two complex backend features needed to move from requirements through architecture, implementation, and independent verification without allowing AI agents to rewrite scope or promote environments on their own.

Action / Outcome

Applied the governed workflow to complete Chainlink VRF-based verifiable dealing and match history. The VRF flow had already passed Base Sepolia request-to-fulfillment validation; both feature sets were merged and deployed to dev for integration testing on 2026-08-24. Neither is deployed to production.

SystemsGoPostgreSQLNATS JetStreamChainlink VRFIndependent Verification
OPS-06INTEGRATED

Unity Game SDK Architecture

IN PRODUCTWebSocket module integrated

Mission / Challenge

Unity developers needed a consistent client structure for product WebSocket integration.

Action / Outcome

Architected the C# Game SDK and WebSocket module, supplied the core implementation example, and supported integration of the module into the product.

SystemsC#UnityWebSocketSDK Architecture

EXPERIENCE

/

MISSION LOG // 03

LOG-2024Taipei, Taiwan

Mega Star Technology Co., Ltd.（萬星科技）

Senior Backend Developer

Functional role: Backend Lead

  • Direct architecture for a 39-service Go platform and GCP/GKE operations across 370+ development, staging, and production pods as of July 2026.
  • Raised validated single-tournament capacity from approximately 150 to 1,000 simultaneous players across 100+ tables through database read/write separation, concurrent processing, Redis caching, NATS storage-I/O tuning, and sustained WebSocket load testing.
  • Integrated three payment providers with signature verification, Redis-backed idempotency, transaction-state controls, callback-order handling, and reconciliation support.
  • Built and led a four-person backend team, directly managing three engineers and mentoring three additional engineers over time; established sprint, architecture-review, code-review, QA, diagram-first, and event-driven practices.
  • Reduced typical API-document drafting and review time from approximately 30 minutes to under five and introduced fixed AI-assisted workflows across product management and marketing.
  • Designed and operate Digital Forge, a human-governed AI engineering system used across requirements, implementation, independent verification, release, maintenance, database exports, and SRE work; applied it to complete two backend feature tracks that entered dev integration testing on 2026-08-24.
LOG-2022Taipei, Taiwan

9dimensions（玖緒）

Senior Backend Engineer

  • Assumed primary architecture responsibility in February 2023 for the 39-service Go platform, defining service boundaries, data flows, NATS subjects and event contracts, cache strategies, repository structure, and production architecture.
  • Translated product requirements for the real-time game engine and operations dashboard into service architecture, implementation plans, and client-server contracts.
  • Built a NATS-based real-time chat system and introduced read/write separation to reduce bottlenecks; refactored an encrypted-chat backend with dependency injection and unit testing to improve modularity and reliability.
LOG-2018Taipei / Remote

9dimensions（玖緒）

Senior iOS Engineer

  • Built reusable networking abstractions, generic UI components, and development tools across real-time news, multimedia, chat, discussion, membership, and internal video applications, improving delivery efficiency by an owner-estimated 10%+.
  • Mentored a junior developer and built a membership-points application whose model tests reached 87% coverage.
  • Worked part-time and remotely from Sep 2019 through Dec 2020 while studying full-time in the United States, then returned to full-time employment.
LOG-2020Los Angeles County, California

Pacico, Inc.

iOS Developer Intern

  • Refactored networking code into a dedicated module, released fixes across iOS and Android, and managed TestFlight, App Store, and Google Play release workflows during a CPT internship.
LOG-2015Taipei, Taiwan

Taisol Electronics

Software Engineer

  • Built factory-test iOS and Android applications that automated error-report generation and reduced assembly-line manual testing time by an estimated 25%+.
  • Developed Bluetooth-enabled prototypes to speed buyer-specific customization validation.
LOG-2014New Taipei City, Taiwan

GIGABYTE

Engineer

  • Built a mobile notification system and reusable C# and Java API frameworks with OAuth, enabling designated iOS and Android applications to receive messages from external API servers.
  • Optimized MSSQL and MySQL stored procedures, reducing average query time by an estimated 30%.
LOG-2012New Taipei City, Taiwan

GIGABYTE

Engineer Intern

  • Built internal and foundation websites and created an article-generation tool that reduced estimated publishing time per article by approximately 50%.

APPROACH

/

OPERATING PRINCIPLES // 04

01

Make time for what matters

Use AI to reduce repetition and information overhead, then spend the returned time deliberately on deeper work, helping others, family, and learning.

02

Turn repeated effort into reusable systems

Convert recurring work into structured context, defined workflows, evidence, and reviewable decisions instead of relying on isolated prompts.

03

Learn with AI; keep judgment human

Let AI help organize information, surface questions, and perform defined work while people retain judgment, relationships, sensitive decisions, and accountability.