Boshow — commerce built around creators
Overview
Hi — I'm Taken.
I studied AI in school, then picked up mobile and web so I could actually ship models in products.That mix made me want a portfolio that felt a little less template-y — so I built this site.Today I split my time between systems integration and AI R&D — I'm not glued to every headline, but I stay close to what actually ships.Lately I'm most curious about diffusion models and 3D / generative AI.
Education
14.03 — 17.02
Wondang High School, Incheon
17.03 — 21.08
Daegu Gyeongbuk Institute of Science and Technology (DGIST)
Career
19.01 — 19.02
DGIST Advanced Electronics Devices Research Group
Applied nanoparticles to MoS₂ channels and studied their effects under selective, localized laser annealing.
Follow-on work on selective, localized laser annealing for flexible multilayer MoS₂ thin-film transistors.
19.06 — 19.08
nepes (Co., Ltd.)
Validated KokoaFab classroom circuit kits; built and demoed a companion robotic arm.
20.01 — 20.02 · 21.01 — 21.02
DGIST Intelligent Digital Systems Lab
Project 1 (Jan–Feb 2020): Boosted model quality with LSTM backbones and ensemble methods.
Project 2 (Jan–Feb 2021): Improved a ResNet setup with quantization-aware training and hyperparameter tuning, borrowing ideas from DARTS: Differentiable Architecture Search (ICLR 2019).
20.06 — 21.06
HygenPower (Co., Ltd.)
21.06 — 21.12
Univa
Grand-Star Platform AI Championship — Daekyo track: 1st place & Minister of SMEs and Startups Award (Nov 2021).
22.01 — 22.11
Oncloud9 (Co., Ltd.)
Systems integration (SI) developer
22.11 —
Zefit (Co., Ltd.)
AI & SI — professional researcher (alternative service)
Introduction
• Default to being careful — I design and build with the full solution space in mind.
On every project I aim for clear performance targets — latency, accuracy, whatever matters. If I spot a better approach mid-flight, I dig in instead of deferring it. When a technical choice ripples into product scope, I pull planners in early and reset direction together.
• I share context as I go.
New tricks I pick up — a faster debug workflow, a cleaner ML recipe, a publishing gotcha — I pass along so the whole team ships faster across AI, web, and spec work.
Skills
Frontend
Primary
ReactNext.jsVueReact NativeNode.js
Also use
Flutter
Backend
Primary
NestJSExpressMySQLPostgreSQLStrapiRESTFastAPI
Also use
Vanilla JSFlask
AI & tooling
ML & small desktop utilities
PyTorchTensorFlowC++PyQt
Infrastructure & cloud
AWSNaver CloudEliceSaladDocker
Work
Projects
Selected shipped work — separate from the résumé sections above.
Avantgarde × SuperNormal — landing page
Goodnews — Catholic rosary app
Quiztok — major app refresh
