01
LLM Products & Agents
Conversational AI, retrieval, and autonomous agents wired into real products — from prompt design to evals and human-in-the-loop guardrails that survive production.
LLMsAgentsRAGAI SDK
Studio
EddieLabs / 에디랩스
EddieLabs / 에디랩스
Discipline
AI · Web · Systems
AI · Web · Systems
Status
● Open for work
● Open for work
Based in
Seoul · KST
Seoul · KST
[ AI & software studio — Seoul ]
We ship AI
& software with
precision.
EddieLabs is a founder-led studio by Eddie — an NVIDIA-certified AI engineer. We build LLM-powered products and durable web software: models in production, interfaces that hold up.
system status
liveseoul · kst––:––:––
p95 latency● healthy
- runtime
- edge · cuda
- region
- icn1 · seoul
- stack
- llm · next · ts
TypeScript
Next.js
React
Node
Postgres
Design Systems
Motion
Edge
AI / LLM
Performance
TypeScript
Next.js
React
Node
Postgres
Design Systems
Motion
Edge
AI / LLM
Performance
[ 01 / Services ]
What we do
Four disciplines, one standard. Every engagement is handled end-to-end with the same obsessive attention to detail.
02
Applied ML & GPU Compute
Diffusion models, transformer-based NLP, and anomaly detection — trained and accelerated with CUDA across single or multi-GPU setups. NVIDIA-certified throughout.
PyTorchCUDADiffusionNLP
03
Web & Product Engineering
Full-stack applications on Next.js and TypeScript — from marketing sites to complex dashboards, architected to scale and ship on schedule.
Next.jsReactTypeScriptPostgres
04
Interfaces & Performance
Precise, accessible interfaces backed by reusable component systems, deployed edge-first and measured relentlessly. We cut the milliseconds that matter.
Design SystemsAccessibilityEdgeVercel
[ 02 / Selected Work ]
Things we've shipped
A selection of builds — from LLM projects to interface work.
[ Coming soon ]
Projects will be listed here shortly.
[ 03 / About ]
Led by Eddie — an engineer who cares about the details.
EddieLabs (에디랩스) is an independent AI & software studio in Seoul, led by Eddie (Joungho An). NVIDIA-certified across generative AI, transformer NLP, and CUDA — with a BA in Digital Media, Culture & Tech from Royal Holloway, University of London — I partner with founders to ship AI products that feel inevitable: fast, precise, and built to last.
8×
NVIDIA AI certs
19+
Open-source repos
EN · KR
Bilingual delivery
1:1
Founder-led builds
How we operate
01
Precision over noise
Every prompt, pixel, and millisecond is deliberate. We remove more than we add.
02
Ship, then sharpen
Real software in production beats perfect software in a doc. We move fast and refine relentlessly.
03
Model to interface
From training and CUDA acceleration to the UI users touch, we take end-to-end responsibility for the outcome.
[ 04 / Contact ]
Let's
build it.
Tell us what you're making. We reply to every serious inquiry within one business day.