About Brandon

I came from film, stayed for systems, and now build toward AI infrastructure.

Full-stack AI engineer at Playtag in Seoul, working on mono_xyz, AI video analytics for human behavior, across product, backend, and infrastructure.

I started in film and theatre, moved into software in 2019, and grew through frontend, full-stack, backend, DevOps, and AI-native engineering work.

The through-line is not a title. It is the way I work: trace the system, name the tradeoffs, verify with tests and logs, then write down what changed my mind.

Korean / English
Seoul
calendar events synced at MOBA
7M
MOBA sync throughput gains
5-20x
software switch from film
2019

Career arc

  1. 2004-2012

    Film and theatre at Hanyang

    Studied cinematography, film directing, sound, and editing before software became the craft.

  2. 2019-2021

    Self-study into engineering

    Built a foundation through JavaScript, Java, C, Nand2Tetris, algorithms, and production-oriented web work.

  3. 2021-2025

    Frontend to backend ownership

    Worked across MODULABS, Moviation, and Playtag, moving from product UI to full-stack delivery, service migration, and production operations.

  4. 2025-2026

    MOBA backend, DevOps, and AI data systems

    Sole-maintained calendar sync while owning billing surfaces, real-time updates, AWS/Terraform infrastructure, and Airflow-based data pipelines.

  5. 2026-now

    Back at Playtag as a full-stack AI engineer

    Returned to Playtag on August 31, 2026 to work on mono_xyz, AI video analytics for human behavior, across product, backend, and infrastructure.

What I build now

  • AI video analytics at Playtag

    mono_xyz turns everyday footage into person-level behavioral insight; I work on it full-stack, across product, backend, and infrastructure.

  • Infrastructure that the team can operate

    Terraform, CI/CD, and AWS deployment paths that are reviewable, repeatable, and fail visibly instead of silently.

  • Agent-native engineering

    B3, reviewer loops, and verification tooling that keep AI-assisted work auditable, from the first prompt to the merged diff.

Projects

Systems I build and maintain outside the day job.

  • Brandon's Binary Brain

    B3

    A version-controlled personal operating system for notes, agent rules, skills, reviewer loops, and decision memory.

  • AI knowledge platform

    Crucio

    A personal AI system for model experimentation, knowledge workflows, and pipeline inspection.

  • public learning surface

    brandonwie.dev

    The place where production notes become essays after the evidence is strong enough to share.

Operating principles

  • Correctness before cleverness

    Timezones, recurrence rules, soft deletes, orphan rows, and race conditions deserve boring precision.

  • Evidence beats confidence

    I use AI heavily, but decisions still need tests, logs, diffs, production data, and explicit ownership.

  • Write the tradeoff down

    Good systems work is remembering why a decision was reasonable when the context is no longer fresh.

Learning edge

Building toward

MLOps and AI engineering, without pretending the work is already done.

The next edge is deeper computer science and AI systems work: graduate-level foundations, production ML infrastructure, and tools that make agents easier to verify.

  • Georgia Tech OMSCS target: Spring 2027
  • GTx math and algorithms certificates
  • AWS Developer and Solutions Architect track
  • Operating systems, concurrency, and lower-level systems depth

Contact

The best ways to reach me.