AI Specialist · Digital Government · International Development

Hoon Jung

From Python to Policy.
Connecting AI technology and policy.

Two decades bridging hands-on AI delivery and public-sector strategy — from software engineering, through digital-government and ODA consulting across 15+ countries, to AI project delivery and responsible-AI research. Currently an Independent AI Advisor & Researcher.

20
years in technology
68
projects delivered
25
AI projects
15+
countries
01The core

AI, from the technical to the political

AI approached not from the model outward, but from the institutional and regulatory context inward — delivery and governance as one practice.

In practice — Python

AI Technical Delivery

Leading AI projects end-to-end across healthcare, telecom, finance, and the public sector — ML/DL, LLM fine-tuning, and GenAI agents on cloud-native infrastructure.

ML / DLLLM · LoRAGenAI agentsMLOps
Representative work
  • Surgical-complication prediction AI + CDSS — Ajou University Hospital · 85% detection accuracy
  • Brain-disease detection platform on domestic NPU — KT Cloud / Rebellions · CNN + Llama 3.1
  • Telecom GenAI agent, Uzbekistan — Turon Telecom · LangGraph + RAG
  • Automated LLM evaluation system — ETRI
  • Enterprise MLOps & GenAI platform — KMAC
In research — Policy

AI Governance Research

Responsible-AI research on safety, alignment, and governance — designing guardrails at the model level while advising ministries on AI policy in the same engagement.

GuardrailsAlignmentBiasGovernance
Focus areas
  • Bias detection in LLM outputs — self-evolving LLMOps (national R&D, PI)
  • Responsible-AI governance architecture — national R&D planning
  • AI ecosystem & education policy — feasibility studies for governments
Supported by
02The foundation

Grounded in reach and depth

Two axes of experience underpin the AI work — a global footprint, and deep public-sector consulting.

Global

International cooperation · Global business
Former Consultant · OECD (Paris, 3 yrs)

30 ODA engagements across 15+ countries — feasibility studies, capacity building, and government advisory, including a decade of work in Uzbekistan.

Representative work
  • Uzbekistan startup ecosystem — 2-yr FS leading to $5M KOICA-funded U-ENTER
  • e-Constitutional Court modernization — Uzbekistan
  • National AI ecosystem & education FS — Uzbekistan & Jordan
  • Global Mobile Innovation Project — OECD HQ, Paris
  • Integrated complaint system — Indonesia–UNDP
15+ countries · Uzbekistan, Indonesia, Tunisia & more

Consulting

Digital Government · Informatization Strategy

Digital government, data governance, and informatization strategy for public institutions and major enterprises — turning digital strategy into working systems.

Representative work
  • Cross-agency data governance, 7 institutions — Tunisia (G4B)
  • 10-country public-data readiness diagnostic — ASEAN · 70+ indicators
  • Digital tax administration with fraud detection — Costa Rica (HomeTax model)
  • Maritime big-data strategy (ISP) — Korea Coast Guard
  • Cloud modernization (K8s, CI/CD) — Mirae Asset, Shinhan Card, DB Insurance
Public sector & enterprise · since 2015
03Research

Responsible-AI research

Safety, alignment, and governance — submitted to flagship venues in AI. Follow a link for details.

DCW

Defensive Context Weaponization

When AI safety guardrails turn a user's own personal context against them.

flagship venue · under review
LJB

LLM-as-Judge Bias

Identity-conditioned defensiveness when an LLM judges criticism of its own work.

flagship venue · under review
MPSG

Multi-Policy Switching Guardrail

A multi-policy safety guardrail framework for large language models.

flagship venue · in revision

Three provisional patents filed. Multiple additional AI research projects are currently underway.

Published — Regional Brain Volume Changes Across Adulthood. Brain Sciences, 15(10), 1096 (2025). Co-author.

Emerging studies
AEBARMMIBAPEDRMEIFB
04Background

Three threads, two decades

Software

A decade of hands-on software engineering (2003–2009) — full-stack e-government and platform development, and a co-founded sports-media startup.

Consulting

Digital government, data governance, and cloud transformation for public institutions and major enterprises across Korea.

International Cooperation

Three years at OECD headquarters in Paris, then 30 ODA engagements across 15+ countries — including a decade of work in Uzbekistan.

NowIndependent AI Advisor & Researcher
EducationM.Sc. AI, aSSIST
Executive MBA, SDG Management School (Geneva)
B.S. Computer Science & Business, Handong
LanguagesKorean · English · French
Contact

Get in touch

An integrated practice — connecting AI strategy, technical delivery, and governance advisory. For collaboration, research, or advisory conversations, you're welcome to reach out.

nayahun2@gmail.com