About
Chaojie Zhu
AI engineer with 6+ years spanning computer vision, quantitative modeling, domain-adapted LLMs,
and agentic AI systems. Currently the sole AI engineer at Zand, UAE's first AI-native digital bank.
Current Role
Senior Generative AI Engineer · Zand
Best Initiative in Agentic AI of the Year 2026
Middle East Banking AI & Analytics Summit
Middle East Banking AI & Analytics Summit
Sole AI engineer (reporting to AI Lead) responsible for designing, building, and shipping the bank's agentic AI platform end-to-end — owning architecture decisions, model selection, infrastructure, and production deployment.
Key deliverables:
Multi-agent orchestration framework powering onboarding, compliance screening,
AML investigation, and customer support workflows
KYC/KYB onboarding agents — document intake, OCR/extraction, sanctions & PEP
screening, automated risk scoring
AML compliance copilots — case triage, adverse media summarization,
investigator workflows with full audit trails
Multilingual support copilot (RAG + tool use) with safe human escalation,
Arabic and English
Key Project
Masaar · AI-Native KYC/AML Platform
Independently architected within Zand. A full-stack compliance platform where AI agents serve as first-class workflow nodes — users define KYC/KYB verification flows through natural-language conversation. Supports all major verification types with configurable agent pipelines: document extraction, registry checks, sanctions screening, risk scoring, and adverse media analysis.
Career
Journey Timeline
2024 – Now
Zand Digital Bank, Dubai Sole AI Engineer. Built the bank's entire agentic AI platform from scratch.
2023 – 2024
Guangming Lab, Shenzhen Led CarbonGPT — multilingual domain-adapted LLM for carbon markets, 4 languages, 100K+ entries. Pioneered GenAI workflows for financial compliance.
2020 – 2022
WeBank, Shenzhen Quant trading (Sharpe > 2.5), CNN-based HFT with 180+ sub-models, remote sensing CV platform, edge AI on satellite hardware.
2017 – 2020
Beihang University (BUAA) M.Eng in Software Engineering (AI direction)
Open Source
Side Projects
Skills
Technical Stack
Agentic AI
- Multi-agent orchestration
- RAG & tool use
- Workflow engines
- Guardrails & evaluation
- MCP integration
LLM / NLP
- Domain adaptation
- Prompt engineering
- Fine-tuning & alignment
- Constitutional AI
- Embedding models
AI Coding
- Claude Code
- Codex
- OpenClaw
- Vibe coding workflows
- Context engineering
Stack
- Python · TypeScript
- React · Node.js
- FastAPI · Fastify
- PostgreSQL · Redis
- Docker · Azure OpenAI
Domain
- KYC / KYB
- AML compliance
- Sanctions & PEP screening
- UAE CBUAE regulations
- Carbon neutrality
ML / CV
- LightGBM / CNNs
- YOLOv8 detection
- Remote sensing
- Edge AI inference
- Quantitative trading
Writing
What I Write About
Agentic AI in Production — Multi-agent orchestration, guardrails,
evaluation, and shipping AI in regulated industries.
AI-Augmented Development — Claude Code workflows, vibe coding methodology,
anti-sycophancy frameworks, and working with AI coding agents.
Architecture & Engineering — System design for AI-native applications,
technical trade-offs, and building things that survive production.