Lee ChanyangBackend Engineer

Language

Lee ChanyangBackend Engineer · 4+ years · From requirements analysis to operational architecture

I structure domain rules and reliably improve services in production.

AI is a tool; judgment and accountability are mine. I analyze requirements with complex rules and exceptions, design for duplicate processing, data consistency, and failure recovery, and take responsibility for verifying the results myself.

Selected work

Selected Projects

Only cases with a verifiable scope of my own implementation are included. Each detail page covers models, invariants, failure scenarios, and limitations.

01Goodbye BlueGlobalContent Commerce BackendLive

Educational Content & Community Service

I connected video course purchases, small community rewards, and point grants based on offline package payments in a single educational service backend, incorporating language, country, and currency rules into the models.

Company
Korea Institute of Psychological Education
Role
Backend Engineer
Team
BE 1 · FE 2 · Designer 1 · PM/Ops 1
Period
2025.09–2026.05 · 9 months
Responsibility
User/Admin API · DB · Async/Batch · AWS Ops
14M+ pt/monthSettlement volume
0 settlement errors4 months of operation · Based on monthly reconciliation (2026.04–07)
0 duplicate payoutsHorizontal scaling · Idempotency · Row locks
NestJSTypeScriptPostgreSQLPrismaBullMQRedisAWSNICEPAY · PayPal (PG)Claude · Codex
View project
02Video CenterGlobalVideo Payments & SubscriptionsLive

Individual Video Purchases & Subscriptions

This service connects language, currency, Korean and international payment gateways, and viewing access in one backend flow so existing offline customers and online users can buy and subscribe to content in the same video center. By reusing the order, payment, and access models and payment gateway/Outbox patterns validated in a preceding platform, I deployed the video center's payment and subscription backend to production within 3 weeks of receiving the requirements document.

Company
Korea Institute of Psychological Education
Role
Backend Engineer
Team
BE 1 · FE 2 · Designer 1 · PM/Ops 1
Period
June–July 2026 · 3 weeks
Responsibility
User/Admin API · DB · Async/Batch · AWS Ops
3 weeksDeployment using shared models
1 per billing cycleDuplicate-charge protection
NICEPAY · PayletterKorean and International Payment Gateway Boundaries
NestJSTypeScriptPostgreSQLPrismaBullMQRedisOutboxAWSNICEPAY · Payletter (PG)Claude · Codex
View project
03PoET platformData Pipeline

On-Chain Activity Rewards & Statistics Data Pipeline

I connected contract approval, on-chain event collection, reward statistics, and replayable batches into a single data flow.

Company
Algorith Capital
Role
Backend Engineer
Team
BE 5 · FE 2
Period
October 2023–March 2024 · Approximately 9 weeks
Responsibility
User/Admin API · DB · Data Pipeline/Batch · Azure Ops
670 million+Source transaction volume
10,000Search window size
50Maximum processing batch size
NestJSTypeScriptPostgreSQLMySQLTypeORMRedisAzureethers.jsChatGPT
View project

Additional project · Research prototype

Additional Projects

Research prototypes that tested unfamiliar hypotheses with data and developed them into repeatable workflows.

Hypothesis Testing & Data Pipelines

Algorith Capital

Kimchi Premium Research

Kimchi Premium Threshold Strategy Validation PipelinePrototype

Backend / Data Prototype

I explored buy and sell thresholds using hourly price data, compared performance metrics, and extended the work into a pipeline that updates data hourly.

  • 3 years, 8 months · Hourly validation data
  • 0.5 increments · Buy and sell threshold search
Explore this project

Personal project · How I work

Personal Projects

Personal projects exploring problem definition, experimental design, implementation verification, and publicly shareable limitations outside production services.

Experience Data & Information DesignLive

Personal Projects

Evidence-Based Engineering PortfolioIn Progress

Portfolio Designer / Frontend Engineer

An engineering portfolio that began as a quick visual draft and evolved as I identified reading friction, distance from supporting evidence, and public-disclosure boundaries.

  • 8 · Project detail pages built on a shared schema
  • 13 PDRs · Append-only decision records
Explore this project
Screen Perception & Automation Runtime

Personal Research Project

Screen-Recognition Automation EngineIn Progress

Automation Systems Engineer

I am developing a personal automation-runtime project that recognizes objects and states in screen environments that are difficult to control through APIs or selectors, executing actions only in verified states.

  • 13 · Target-specific detector directories
  • 17 · Profile-based execution variants
Explore this project
AI System Design & Research into Agent Limitations

Personal Research Project · On hold since the final commit in 2026-06 (c27a98f) · Observer = the operator (n=1)

Self-Improving Multi-Agent HarnessPaused

System Designer / Engineer (sole responsibility for design, implementation, and operation)

While building a personal multi-agent harness capable of self-improvement, I used actual operational data to investigate which agent limitations engineering can reduce, which are irreducible, and how cognitive tasks should be allocated across models.

  • 15%→60% · System functioning rate after connecting memory to the runtime
  • 71%↓ · Base-load tokens (brain-distill, measured with 4-core)
Explore this project

Production experience

Production Backend Experience

I have worked on service reliability across initial development, scaling, changes, and recovery.

01

Build

I implemented complex service rules as data structures and processing flows that could run in production.

  • Separated scheduled points payouts, finalized history, and current balances — 187 verification tests
  • Separated orders, billing cycles, payment attempts, and authorization results — 527 payment-path tests
  • NICEPAY · Payletter one-time payments and subscription lifecycle — 168 subscription tests
  • Settlement batches and ledger consistency — 87 settlement tests

Domain models · DB constraints · 2,973 regression and integration tests (969 covering core paths)

Domain implementation evidence
02

Scale

I incrementally scanned a source table with 670M+ rows and converted it into rewards statistics.

  • Search windows of 10,000 IDs from the last processed position
  • Process 50 records at a time and resume the next range after one second
  • Save statistics results and the last processed position together

Batch code · Cursors · DB transactions

Large-scale batch-processing evidence
03

Change

I safely transitioned deployment, traffic, and runtime boundaries while the service was running.

  • GitHub Actions → S3 → CodeDeploy artifact path
  • Shared ALB Host Header routing → Service-specific ALB
  • Single EC2 instance → ASG and Launch Template operating criteria

Infrastructure documents · Deployment records · Runbooks

Traffic-boundary transition evidence
04

Recover

I used reprocessing to bring state back into agreement even when external operations succeeded and internal operations failed.

  • Recover Redis and BullMQ failures from database state
  • Guard against duplicate charges when database finalization fails after gateway authorization
  • Release stale batch claims and isolate individual failures
  • Verification spec files: Outbox 44 · Idempotency keys 23 · Reprocessing 23 · Row locking 3

State machines · Idempotency keys · Outbox reprocessing verification specs

Recovery Design Evidence

Career

I have analyzed requirements and designed, implemented, and operated backends across different service domains

Most recently, I owned payment, subscription, and points backends and operational infrastructure for education, community, and video services in production. Previously, I worked on on-chain data pipelines, content SaaS, and legacy modernization.

  1. Korea Institute of Psychological Education

    Backend Engineer

    Solely responsible for building and operating payment, subscription, point, and notification backends, databases, and AWS infrastructure for two production services in education/community and video

  2. Algorith Capital

    Backend Engineer

    On-chain reward and statistics batches tracking a source table with 670M+ rows, contract approval workflows, and strategy backtesting

  3. ToSpace

    Backend Engineer

    Responsible for backend APIs for location-based AR gift commerce and content creation/distribution SaaS, asynchronous 3D file processing, and AWS/ECS deployment architecture

  4. Life Planning Lab

    Backend Engineer

    Responsible for goal/content APIs, databases, data migration, and renewal of authentication and AWS operations foundations for two services, Attale and Attale Pro

Capabilities

Technology Stack

Technologies used in selected projects, organized by role.

Core Backend

  • TypeScript
  • Node.js
  • NestJS
  • REST API
  • Swagger

Data & Messaging

  • PostgreSQL
  • MySQL
  • Prisma
  • TypeORM
  • Redis
  • BullMQ

Cloud & Delivery

  • AWS
  • Docker
  • GitHub Actions

Payments & Notifications

  • NICEPAY
  • Payletter
  • PayPal
  • FCM

Frontend

  • React
  • Next.js