JL

Ji-won L.

Vetted Crafter

Staff Software Engineer – MVP & Deployment

🇰🇷South Korea·8 years experience
Available in 1 monthFull-time

I build MVPs and deploy them — from concept to production in weeks.

About

I build MVPs and deploy them. I have shipped 5 products from concept to production in 4–8 weeks. My focus is on getting something live fast: Next.js, Vercel, Docker, GitHub Actions. Cursor helps me scaffold apps and write deployment configs. I believe that deployment is part of the product — if you cannot ship it, it does not exist.

I work with TypeScript, Next.js, Docker, and PostgreSQL. I have set up CI/CD for startups, migrated from manual to automated deploys, and built zero-downtime release processes. I deliver MVPs that are production-ready from day one.

AI Expertise

MvpDeployment
TypeScriptNext.jsDockerVercelGitHub ActionsPostgreSQLCursor

Notable Projects

Smart Factory Worker Safety System

Built a real-time worker safety monitoring system for a Korean automotive parts manufacturer. Uses YOLOv8 for PPE detection (helmets, vests, gloves) and ByteTrack for worker tracking across 40 camera zones. Alerts trigger in under 300ms; dashboard built with React and FastAPI WebSocket streaming.

YOLOv8PyTorchOpenCVByteTrackFastAPIReactTensorRT

Workplace safety incidents reduced by 54% in first year; insurance premiums reduced by $180K annually; system passed Korean industrial safety certification.

Multimodal Product Quality Report Generator

Built an application that combines YOLOv8 defect detection on product images with GPT-4 Vision for contextual defect description generation. The system identifies defects, classifies severity, and uses the LLM to write human-readable quality inspection reports in both Korean and English.

YOLOv8GPT-4 VisionOpenAI APIPyTorchFastAPIReact

Report generation time reduced from 45 minutes to 3 minutes per inspection; defect documentation quality rated higher by QA auditors than manual reports in blind comparison.

Retail Crowd Analytics Platform

Designed a full-stack retail analytics product tracking customer flow, dwell times, and zone heatmaps across department store floors. Built the vision pipeline on YOLOv8 + DeepSORT, deployed on Vertex AI, with a React dashboard and BigQuery data warehouse for historical trend analysis.

YOLOv8DeepSORTVertex AIBigQueryReactFastAPIPython

Deployed across 12 department stores; store layout optimization guided by analytics data increased revenue per square meter by 11% on average.

Work Experience

Staff Computer Vision Engineer

Kakao

2021 – Present

Lead vision AI for Kakao's commerce and AR features serving 50M+ Koreans. Own the real-time object detection infrastructure and the multimodal AI product strategy.

Computer Vision Engineer

Samsung SDS

2017 – 2021

Built vision AI systems for smart factory applications across Samsung's manufacturing subsidiaries. Delivered defect detection, robotics vision, and worker safety systems in automotive and electronics manufacturing.

Education & Certifications

🎓

M.Sc. Electrical Engineering (Computer Vision)

KAIST · 2016

🏆 NVIDIA Deep Learning Institute – Computer Vision🏆 Google Professional ML Engineer🏆 Google Cloud Professional Data Engineer

Interested in working with Ji-won?

Tell us about your project and we'll facilitate an introduction.