ENGINEER · 03

JungChan LEE

ML SYSTEMS ENGINEER  ·  PRODUCTION PIPELINES

Self-taught ML engineer. Build production Python pipelines for LoRA fine-tuning, parameter optimization, and identity evaluation. Solo lead at The Blue Studio AI Lab. Sole author of the AZIT LoRA pipeline — nine Python scripts, training on RTX 5090, evaluated against a HuggingFace baseline.

ARCFACE BENCHMARK  ·  AZIT FEMALE ACTOR LORA
0.6050 avg similarity
HuggingFace reference: 0.5090  ·  +18.9% over baseline
MOST RECENT ROLE
AI Lab Lead
The Blue Studio — sole engineer
PERIOD March 2026 — April 2026
SCOPE Generative-image LoRA training, inference workflows, pipeline build
STACK Python · ComfyUI · AI-Toolkit · ArcFace
STATUS Open to new roles
EXAMPLES

Working environments

Two captures from the AZIT pipeline build — the dataset preparation view in AI-Toolkit, and the inference graph in ComfyUI. Full step-by-step trace in  VISUAL WORKFLOW ↗.

AI-Toolkit dataset view — multi-view subject portraits prepared for LoRA fine-tuning

AI-Toolkit dataset view — multi-view subject portraits prepared for LoRA fine-tuning

ComfyUI inference workflow — LoRA load, dual CLIP encode, KSampler, FaceDetailer (Impact Pack)

ComfyUI inference workflow — LoRA load, dual CLIP encode, KSampler, FaceDetailer (Impact Pack)

HONEST FRAME

B.A. in Digital Content (2025), self-taught ML engineering through one production project at depth. I do not have years of web-scale backend experience; I do have evidence I can scope, build, ship, and measure a non-trivial ML pipeline solo, with documentation and quantitative outcomes.

I prefer to underclaim and let the artifacts — nine scripts, a benchmark, a structured sweep, a visual workflow trace — do the talking.