하태라ハ テラ

Tera Ha

Game Server & Backend Engineer

With functional programming in Scala 3 and ZIO, I build a stateful distributed game server that runs on event sourcing, CQRS and the actor model. I joined Devsisters a year before launch, and I build the game server for CookieRun: Crumble, a mobile RPG that launched globally in five languages and was played by 2.5 million people in its first month.

Over more than a year before and after launch, I have worked on many layers of the game server myself: content design and implementation, consistency in distributed transactions and event streams, the global launch and live operations, and code quality and team productivity. I build AI into the team's development workflow, but I confirm conclusions with code and evidence.

NowSince the global launch on July 30, 2026, I have handled live updates and operations for CookieRun: Crumble. Launch press release (opens in a new tab)

By the numbers

Public results for CookieRun: Crumble (from company announcements and press coverage)

Players in the first month
2.5M
Inven Global (opens in a new tab)
Cumulative revenue in the first month
₩20B
Inven Global (opens in a new tab)
Free games on the Korean App Store, launch day
#1
Dailian (Korean) (opens in a new tab)
Simultaneous global launch, 2026.07.30
5 languages
Digital Daily (Korean) (opens in a new tab)

Core strengths

  1. Content from design to live

    I owned seasonal boss content, equipment effect rerolls, constellation progression content and an exchange-style shop in one flow, from design through implementation, validation rules and live release. I agree on the rules with the game designers, and I design them so they do not break in the next season or the next update.

  2. I keep distributed systems consistent

    I handle defects that would spread into silent data mismatches along the path through event-sourced actors, two-phase commit, Kafka and CDC. I verified the candidates an AI analysis tool flagged myself, against the code and library source, and fixed them. When a review points out that my judgment was wrong, I look at the evidence again and correct it.

  3. Steady launch and live operations

    Just before launch, I ran AI agents in parallel to audit risks, checked the findings again against the actual code, and caught issues before launch. For every maintenance, I write the work scenario first and get it reviewed before proceeding. I support operations with a kill switch ported from a predecessor project, CS commands I extended, and a live data correction process.

  4. Failures earlier, a faster team

    I moved game design data validation forward from server boot to compile time and the data repository's CI. I analyzed over 2,000 of the team's past review comments with AI agents to rewrite the agent rules, I maintain the AI review bot, and I have reviewed over 740 teammate PRs.

Selected work

Each case study follows the order situation, decisions, result. Internal company figures and non-public details are generalized.

  1. 2026.09 ~ 2026.10Server design, implementation, post-release support

    A season content server, from design to live

    I designed the whole server side of a seasonal boss mode: battles, season settlement, daily rankings and time-limited rewards. After release, I kept handling its live issues.

    • Game systems
    • Event sourcing
    • Rankings
    Read case study
  2. 2026.04 ~ 2026.10Verification, fix scope, failure alerts

    Consistency in distributed transactions and event streams

    I verified candidates flagged by an AI analysis tool and in reviews myself, against the code and library source, decided the scope of each fix, and fixed an ambiguous commit in 2PC and a Kafka rewind that did not work as intended. I found and blocked TTL deletes that nearly flowed into CDC myself, and after launch I checked retry safety across all paths.

    • Distributed transactions
    • Event sourcing
    • Kafka
    • CDC
    Read case study
  3. 2026.06 ~ PresentLaunch audit, performance improvements, maintenance and operations

    Global launch readiness and live operations

    Just before launch, I caught issues with a parallel AI audit, removed waste on the hot path found in load testing, and ported a predecessor project's kill switch, which turns off features without redeploying. After launch, I support operations with maintenance scenarios and reviews, CS commands and a live data correction process.

    • Launch readiness
    • Load testing
    • Live operations
    Read case study
  4. 2026.06 ~ 2026.10Server design, implementation, verification

    Getting probability and numbers exactly right

    In a randomized progression system, I removed probability differences by line position at the design stage, before the feature was released, and kept the numbers on screen from diverging from the server's bonus. Ahead of a higher auto-draw cap, I made the hot path about 2x faster.

    • Game systems
    • Probability
    • Performance
    Read case study
  5. 2025.08 ~ 2026.10Validation rules, compile-time validation, data CI design

    Moving failures to the cheapest stage

    To stop game design data mistakes before they become live incidents, I extended server boot validation and moved validation forward step by step to compile time and the data repository's CI.

    • Data quality
    • Compile-time validation
    • CI
    Read case study
  6. 2025.08 ~ PresentStanding reviewer, running the team's AI development environment

    A team quality gate built on reviews and AI

    I reviewed over 740 teammate PRs and stopped defects before launch, then turned the knowledge built up in reviews into the team's AI agent rules. I investigate together with AI, but confirm conclusions with code and evidence before applying them.

    • Code review
    • AI agents
    • Engineering culture
    Read case study

Experience

2025.07 ~ Present

Devsisters (opens in a new tab)Studio Kingdom

Server Software EngineerCookieRun: Crumble

I developed the game server for CookieRun: Crumble, a new mobile RPG in the CookieRun IP, starting a year before launch. The game launched globally in five languages on July 30, 2026 and was played by 2.5 million people in its first month. I now handle live updates and operations. Over this time, I have merged over 480 PRs and reviewed over 740 teammate PRs.

  • Content: I owned seasonal boss content, equipment effect rerolls, constellation progression content, equipment presets and an exchange-style shop from design to live, and implemented guild operations tools, the gift center, the idle auto-production reward overhaul and the guide quest overhaul.
  • Consistency: Along the path through event-sourced actors, 2PC, Kafka and CDC, I verified in the code and fixed defects that an AI analysis tool flagged or that I found myself, and ran a full sweep to find and close paths that were not safe to retry.
  • Launch and operations: I handled the pre-launch parallel AI risk audit, removing hot path waste found in load testing, porting the kill switch, extending CS commands and live data correction migrations. For every maintenance, I write the work scenario first and get it reviewed before proceeding.
  • Quality and productivity: I moved game design data validation forward to compile time and the data repository's CI, and carried out the Scala 3.8, JDK 25 and ZIO ecosystem upgrades and shortened CI time.
  • AI and collaboration: I analyzed over 2,000 review comments with AI agents to rewrite the team's agent rules, and I maintain the AI review bot a teammate introduced. I work with the game design, QA, data analytics, client and operations teams, and exchange changes in both directions with an in-house predecessor project on the same tech stack.
  • Scala 3
  • ZIO
  • gRPC
  • Protobuf
  • CockroachDB
  • Redis
  • Kafka
2023.07 ~ 2025.05

Vortex Gaming (Inven) (opens in a new tab)

Backend EngineerWeb3 social network & community for gamers

I joined a new-business project as an early member and handled backend design, implementation and deployment as it grew into a service with hundreds of thousands of MAU.

  • Built LLM-based RAG search, summarization and multilingual translation on my own.
  • Built real-time 1:1 and group chat with WebSocket and RabbitMQ.
  • Introduced a CI/CD pipeline, ELK logging and Slack alerts, which stabilized deployments and incident response.
  • Spring Boot
  • FastAPI
  • Python
  • MySQL
  • MongoDB
  • Redis
  • Elasticsearch
  • RabbitMQ
  • Docker
  • Jenkins
  • GitHub Actions

Education

Awards

  • 2023School Hackathon, Top Excellence AwardFallTect: a service that detects falls with AI and sends SMS alerts in real time (Python, FastAPI, OpenPose)
  • 2022School Algorithm Contest (DAPC), Silver Prize
  • 2021School Algorithm Contest (DAPC), Bronze Prize
  • School Top-of-Class Award (Top Excellence) by subjectAlgorithm Practice, Introduction to AI, AI Project Practice, Probability and Statistics, Software Engineering Practice and more
  • 2018.09National Student Science Invention Contest, Minister of SMEs and Startups Award

Activities

  • 2020.01Assistant instructor, advanced Python course, Halla University LINC+ coding camp for gifted students
  • 2018Completed the gifted program in information science, Gifted Education Center for Information Science, Gangwon Education Research and Information Institute
  • 2017Completed the gifted program in mathematics, Gifted Education Center, Gangwon Provincial Office of Education

Certifications

  • 2023Craftsman Information Processing Operation (Korean national certificate)
  • 2021Craftsman Information Processing (Korean national certificate)
  • 2020GTQ (Graphic Technology Qualification) Level 1
  • 2019ITQ Excel, Grade A
  • 2018ITQ PowerPoint, Grade A

Open source

Small contributions where I fixed problems I ran into in libraries I use at work, directly upstream.

Skills

Server

  • Scala 3
  • ZIO
  • Event sourcing
  • CQRS
  • Actor model
  • gRPC
  • Protobuf

Data / messaging

  • CockroachDB
  • Redis
  • Kafka (CDC, Outbox)
  • MySQL

Operations / infrastructure

  • Kubernetes
  • Helm
  • GitHub Actions
  • Datadog
  • Docker

Quality / performance

  • JMH
  • Testcontainers
  • Load testing
  • Scala 3 macros
  • WartRemover

AI-assisted development

  • Claude Code
  • Codex
  • Agent rules (AGENTS.md)
  • AI code review bot maintenance
  • Parallel agent audits and fact-checks

Earlier projects

AI services I built in high school. I ran a booth at a COEX exhibition for all three.

  • 2023.05Solo project

    YeonDalm

    A DeepFake service that finds your celebrity look-alike and turns your face into a celebrity's face.

    I built the server with FastAPI, trained EfficientNet for classification on data I crawled myself, and implemented the face swap with MobileFaceSwap.

    • Ran a booth at AI Expo KOREA 2023 (COEX)
    • FastAPI
    • PyTorch
    • BeautifulSoup
    • EfficientNet
    • MobileFaceSwap
  • 2022.12Team lead

    Neulboram

    A serialized fiction platform that generates novel covers and characters with AI.

    I developed OAuth login and post CRUD with Spring Boot, and built a Stable Diffusion server with FastAPI to generate covers and characters.

    • Ran a booth at SOFTWAVE 2022 (COEX)
    • Spring Boot
    • FastAPI
    • QueryDSL
    • OAuth2
    • Stable Diffusion
  • 2022.04Solo project

    AI NFT

    A service that generates PFP NFT images in real time with DCGAN.

    I trained and tuned a DCGAN model to improve its performance.

    • Ran a booth at AI Expo KOREA 2022 (COEX)
    • Presented the project to the Prime Minister of Korea
    • Python
    • PyTorch
    • DCGAN