About Me
HaRyeomJangAIAgentEngineer·AIServiceEngineer
Hi — I'm HaRyeom Jang, an AI service engineer who designs and operates AI agents. At Plateer I build the AI agent platform XGEN, covering everything from LLM training and serving to the agent harness and the RAG document pipeline.
6
Research Papers
13
Awards & Honors
8+
Open Source
1
Book (reviewed)
Experience
Career
Plateer Inc.
Oct 2024 – presentAI R&D — Product Development Part Lead
I lead the design, development and operation of the XGEN AI agent platform. I own nearly every layer required to build and run agents — LLM training, serving, the agent harness, the RAG document pipeline — and I present and deliver these projects to customers directly.
- LLM training: SFT/DPO/GRPO and LoRA lineups for POLAR and Gemma, plus the Korean ModernBERT embedding family
- Inference: building and running multi-model vLLM serving with an OpenAI-compatible API
- Agent platform: owner of the XGEN harness, Skills, hybrid memory and MCP-first architecture
- Document → context: designed the 80+ format in-house parser (now the open-source Contextifier) with table- and code-preserving chunking
I design and operate AI agent platforms, and turn what production teaches into libraries and tools anyone can reuse.
Core Competencies
Areas of Expertise
LLM Training Operations
Operating the full training cycle
POLAR·Gemma
lineup
SFT·DPO·GRPO
tuning
ModernBERT
embed
Inference / Serving
Serving and real-time inference optimization
vLLM
serving
OpenAI-compat
api
OmniVoice
realtime
Agent Architecture
Harness · Skills · Memory
Staged
harness
5 LLM
providers
Hybrid
memory
Document → Context (RAG)
Document parsing & RAG pipeline
80+
formats
5 Vision-LLM
ocr
Structure preserved
chunk
Technical Skills
Tech Stack
What I can do and what I have experienced
LLM & Training
Agent & Serving
RAG & Data
Full-Stack & Infra
Journey
Timeline
Enterprise GenAI rollout & technical review of an AI book
Plateer Inc. · Youngjin.com
Launched Lotte Homeshopping's agentic AI service on XGEN, and am currently building Jeju Bank's GenAI platform as part of a KT consortium. I rewrote and technically reviewed the LangChain and LangGraph chapters of "Building AI Services & Agents with LangChain" (Youngjin.com).
Commercializing XGEN & running the Geny research platform
Plateer AI R&D
XGEN started as a personal side project and grew into the company's flagship AI agent platform. I designed the node-edge canvas, the agent harness, the RAG document pipeline, and the MCP/Skills integration. In parallel I run Geny, a VTuber-based multi-agent and real-time TTS research platform where I prototype the next generation of agent components.
Joined Plateer · Lead for AI agent platform development
AI R&D — Product Development Part Lead (Oct 2024–)
I design and operate the full AI agent stack hands-on — LLM training operations, inference and serving, agent architecture, and document-to-context (RAG) — and deliver customer projects on top of it. In August of the same year I completed my M.S. in Big Data Application at Kyung Hee University (GPA 4.075/4.3).
Graduate research · publications and awards
Kyung Hee University Graduate School (academic foundation)
Published in IEEE Access and the Journal of Intelligence and Information Systems, among others. Received the Minister of Trade, Industry and Energy Award twice (grand prize, public-data BI competition) and best-paper awards from the Korea Intelligent Information Systems Society and the Korea Society of Management Information Systems. My research covered GraphSAGE citation-network analysis, renewable-energy virtual power plants, and text mining of live commerce.
Started M.S. in Big Data Application, Kyung Hee University
Kyung Hee University Graduate School
This is where I began research in AI, NLP and network science, and where I first built ML pipelines and large-scale data infrastructure — the starting point of a career that runs from research into products.
B.A. in Counseling Psychology, Sahmyook University
Sahmyook University (B.A. Counseling Psychology)
I majored in counseling psychology (GPA 3.89/4.5) and won several grand prizes in reading and creativity competitions during my studies.
Authorship
Authorship & technical review
Research
Publications
“In the vast universe of knowledge, every research is a star illuminating the path to discovery”
Research Trends of Health Recommender Systems (HRS): Applying Citation Network Analysis and GraphSAGE
H. Jang, J. Yoo, S. Yang (2023)
Journal of Intelligence and Information Systems
The Development of an Aggregate Power Resource Configuration Model Based on the Renewable Energy Generation Forecasting System
E. Kang, H. Jang, S. Yang, S. Yang (2023)
Journal of Intelligence and Information Systems
Exploring Consumer Behavior in the Travel Experience Sector Through Real-Time Live Commerce Chat
E. Kang, H. Jang, S. Yang (2024)
Journal of Hospitality Marketing & Management
Leveraging Stock Discussion Forum Posts for Stock Price Predictions: Focusing on the Secondary Battery Sector
J. Yoo, H. Jang, E. Kang, S. Yang, Y. Yoon (2024)
IEEE Access
HormAS: Hotel Review Management Assistant Service
H. Jang, Y. Yoon, S. Yang, J. Park (2024)
Tourism Management
FedKDA: Decentralized Federated Learning Methods Using Knowledge Distillation and Aggregation in Graph
H. Jang (2025)
In Progress
Achievements
Awards & Recognition
13 Total Awards · 8 First Places · 2018–2024
MOTIE Big Data Competition
Power demand forecasting with big data analysis
MOTIE Big Data Competition
Renewable energy big data analysis
KMIS Spring Conference
AI and Management Information Systems
KMIS Spring Integrated Conference
Data Analytics Research
KIISS Spring Conference
Intelligent Information Systems
Youngrim Won ERP Idea Competition
ERP idea for university students
Kyung Hee University BDGSC
Big Data Graduate Student Conference
Korean Management Conference
Convergence Academic Conference
MIS Integrated Academic Conference
Spring Integrated Conference
Sahmyook Univ. ACE+ Book Story
John Rawls — A Theory of Justice
Sahmyook Univ. 100 Books Contest
George Orwell — 1984 (Team)
Sahmyook Univ. Reading Essay
Gerd Brantenberg — Egalia's Daughters
3rd Addiction Creativity Contest
Creative solutions for addiction prevention
Education
Academic Background
Kyung Hee University
M.S. in Big Data Application
2022.09 – 2024.08 · GPA 4.075 / 4.3
Research in AI, NLP and network science. Multiple journal publications, a best-paper award, and the Minister of Trade, Industry and Energy Award.
Sahmyook University
B.A. in Counseling Psychology
2016.03 – 2020.02 · GPA 3.89 / 4.5
Majored in counseling psychology, winning several grand prizes in reading and creativity competitions along the way.
Connect
Let's Build Together
If you are interested in AI agents, LLM serving, RAG pipelines, or want to talk about a project, feel free to reach out any time.