Looking for a job

Yongjae Kim

@yongjaekim

Message

Product Engineer building production AI tutoring products with hybrid RAG and frontend-first full-stack craft.

South Korea

Message

What I'm looking for

I want to build production AI products where evaluation and accuracy matter—hybrid RAG, robust CI/CD, and measurable cost/quality wins. I enjoy frontend-to-backend ownership and teams that value testing, iteration speed, and clean UX.

I’m a Product Engineer with 4 years of production SaaS frontend experience, and I’m now building full-stack AI products end-to-end. I care about shipping reliably—accuracy, evaluation, and measurable improvements come before demos.

At Nativ, I built a production-grade AI tutor for language learners when fine-tuning wasn’t an option. I implemented a hybrid RAG pipeline (pgvector semantic + tsvector keyword fused with RRF) so the AI answers from the uploaded textbook content, then migrated from Anthropic to OpenAI with per-task model sizing and A/B tested quality.

I also put quality gates into the workflow: I shipped 163 backend integration tests in CI/CD and added an LLM-as-Judge evaluation (GPT tutor → Claude judge) to catch RAG regressions on every prompt change. On top of that, I owned deployment with LemonSqueezy subscription billing and Docker-based CI/CD on low-cost infrastructure.

Previously, I scaled frontend delivery at Bold9 and built advanced UI systems—real-time operations dashboards, high-performance tables, and large form workflows—while raising test coverage from 10% to 70% and cutting CI runtime by 50%. Before engineering, I spent 2 years in international technical sales, collaborating across Asia, ME, and Africa and strengthening my ability to translate user needs into technical execution.

Experience

Work history, roles, and key accomplishments

Current

Full Stack Engineer

Current

Nativ

Feb 2026 - Present (2 months)

Built Nativ, a production-grade PDF-based AI tutor for language learners with a hybrid RAG pipeline that grounds answers in uploaded textbook content. Migrated from Anthropic to OpenAI, A/B tested quality, reduced per-message cost from $0.006 to $0.004 (94% monthly savings), and shipped CI tests plus LLM-as-Judge evals to prevent RAG regressions.

Hybrid RAG Pgvector Full Text Search (Tsvector Ts_rank)LLM As Judge OpenAI API fastAPI Docker React Python

Frontend Engineer

VNTG

Nov 2025 - Feb 2026 (3 months)

Designed a roadmap to automate frontend code generation from Figma designs and requirement documents. Built a 4-stage Spec-to-React pipeline using Claude Code sub-agents, improving HTML quality from 50–70% to 95% and achieving 90%+ Figma-to-React component mapping accuracy with WCAG AA and semantic HTML quality gates.

Figma To React Mapping Spec To React Pipeline Claude Code Quality Gates React Generative AI

Frontend Engineer

Bold9 Ltd

Jun 2021 - Jun 2025 (4 years)

Enhanced a SaaS e-commerce fulfillment platform by improving core pages with advanced search filters and interactive data tables supporting inline editing, bulk operations, and multi-column sorting for smooth 60fps scrolling. Built a real-time operations dashboard and a scalable form system (70+ fields) using Formik/Yup, and increased test coverage from 10% to 70%.

React Dashboard Development Next.js GraphQL Apollo Client Formik Vitest Storybook GitHub Actions Performance Optimization

Overseas Sales Executive

DRGEM Ltd

Jan 2018 - Feb 2020 (2 years 1 month)

Supported global sales of advanced digital X-ray systems by building relationships with Fujifilm branches across Asia, the Middle East, and Africa and providing technical support as a key partner. Collaborated with R&D, quality, and legal teams to resolve technical issues and revise OEM contracts, while gathering market insights through major medical exhibitions (RSNA, ECR, ARAB HEALTH).

Technical Support Cross Functional Stakeholder Management Market Research Sales