marinkim.xyz

Daejeong Kim

Tech Lead · Backend Engineer · AI Product Systems

marin.backend@gmail.com · github.com/Dae-Jeong · marinkim.xyz

I’m Daejeong Kim, a maker who turns possibilities into opportunities and delivers value through products.

Experience

Working language: Korean

MediSolve AI

Apr 2025 - Present

Tech Lead · Backend Engineer | Full-time, early member

  • Proposed Thready and led product design, development, launch, and operation, working with planning, QA, and marketing teammates to build a paid AI content service for creators expanding to Threads.
  • Built the backend, AI generation and evaluation systems, and core user and admin workflows. Used human-rated examples to guide generation and recorded LLM evaluation history for review.
  • Moved generation to background processing after observing users prepare content in advance. Built an agent prototype exposing existing APIs through MCP and began customer A/B testing of conversational workflows.
  • Separated AI execution into its own service and database, retaining product records in the main backend. Used a transactional outbox, retries, and version checks to protect newer state from duplicate or out-of-order delivery.
  • Implemented scheduled drafts, approval notifications, and transactional approval commands. Designed recovery to distinguish pre-call failures from unknown external outcomes and resume the existing publishing attempt.
  • Led planning, design, and development of a medical knowledge base used in internal consultations. Connected terminology curated by clinic operations teams to help correct medical terms in speech-to-text and translation.
  • Led inventory and ordering backends, Redis-based SSO session redesign, and multi-clinic reservation APIs and policies. Co-led the real-time AI consultation backend and implemented operational integrations.
  • Owned operations of the company’s Azure infrastructure, managed with Terraform. Consolidated service instances, contributing to an approximately 30–40% reduction in monthly server costs.
  • Built a standard FastAPI template to help planning, QA, and design teammates implement features with coding agents. Led product planning, issue tracking, and releases, and helped design AI-assisted company workflows.

TheDayLabs

Feb 2025 - Apr 2025

Backend Engineer | Freelance, pre-incorporation engagement

  • Built the initial dermatology CRM backend and established repository structure, development environments, documentation, and issue and pull-request conventions before MediSolve AI’s incorporation. Joined MediSolve AI full-time in April 2025.

Memento AI

Oct 2024 - Jan 2025

Backend Engineer | Intern, then full-time; company later closed

  • Led the manual-capture prepayment integration and added cancellation/refund handling for failed reservations. Moved loyalty-point restoration and pass removal to refund completion.
  • Added immediate and scheduled message/email delivery to an existing multilingual follow-up survey flow, with cancellation, rescheduling, and delivery history.

STUDIO LAB

Dec 2021 - Jan 2024

Vision AI Engineer / Product Manager / Backend Engineer | Full-time

  • Vision AI Engineer (Dec 2021 - Sep 2022): Contributed to apparel-image analysis model development for SellerCanvas.
  • Product Manager (Oct 2022 - Sep 2023): Led SellerCanvas from prototype to v1.0, defining product workflows, feature scope, and release priorities.
  • Translated business requirements into technical specifications for enterprise fashion-brand PoCs, coordinating scope, schedules, and technical validation.
  • Redesigned the product-page creation workflow, defined product-data criteria with fashion specialists, and designed the page-layout algorithm, contributing as a co-inventor of METHOD FOR DISPLAYING PAGE.
  • Backend Engineer (Oct 2023 - Jan 2024): Transitioned into backend development.

Izsol

Aug 2020 - Jun 2021

Vision AI Engineer | Intern

  • Helped establish the student startup Izsol and developed Kidsly, a facial-recognition attendance system. Acquired training data through a national program and trained YOLO models.

Projects

TellingMe

Jan 2024 - Dec 2024

Backend & Infrastructure Lead | Team side project

  • Led Spring Boot backend development and AWS deployment and monitoring. Used Java, Spring, and Spring Boot across three side projects.

Technical Skills

Backend: Python, FastAPI, SQLAlchemy 2.0 async, PostgreSQL, MySQL, Redis, TypeScript, NestJS, RabbitMQ, TaskIQ

AI / LLM: LangGraph, RAG, MCP, LLM integration and evaluation, speech-to-text

Frontend: Next.js, TypeScript (core user and admin workflows)

Engineering Tools: Claude Code, Codex, pytest, Ruff, Pyright, Docker, GitHub Actions, Sentry, Jira

Additional: Java, Spring, Spring Boot (side projects); Azure, Terraform (infrastructure operations); AWS, Vercel (deployment and environment setup)

Military Service

Republic of Korea Army · III Corps

Aug 2017 - May 2019

Radio Communications Operator | Discharged as Sergeant

  • Set up, operated, and maintained military radio communications infrastructure. Assigned wartime duties included antenna installation and communications-node operation to establish and sustain military communications networks.
  • Completed full-term military service.

Education

Woosong University

Mar 2016 - Aug 2021

Game Multimedia major | Graduated

Patents, Awards & Certifications

  • Registered patent: METHOD FOR DISPLAYING PAGE, Korean registration No. 10-2898273. Contributed to the product-page creation workflow design.
  • CES 2024 Best of Innovation, Artificial Intelligence: Contributed to the award-winning SellerCanvas product.
  • KCL AI accuracy certification: Contributed to a product that passed certification by Korea Conformity Laboratories.
  • ADsP: Advanced Data Analytics Semi-Professional, September 2021.

Additional Activities

  • OSS Developer Partnership (Sep 2026): Received and accepted an invitation to an external LLM API platform’s developer partner program based on my public GitHub projects.
  • UX Consulting: Contributed improvement proposals to a team engagement on reward-point visibility and repeat visits. Adopted proposals, together with other product improvements, contributed to App Store ranking and daily-active-user growth.
  • UX Study: Presented improvements to Speak’s post-lesson recommendation flow and was selected as MVP of IPS cohort 12. Also presented a UX-principles analysis of YouTube Music.