Celeste Huang

## Role
Data Security Engineer | AI & Data Security | Security Analytics

## Profile
Security and data practitioner with advanced analytics training, focused on practical security governance systems across data protection, access control, audit analytics, Shadow AI risk control, and AI-assisted security assessment workflows.

## Current Security Work
- Critical financial/fund-system security assessment: map access paths, account permissions, data flows, sensitive operations, log retention, alerting, and control gaps; produce risk findings, remediation suggestions, and evidence-based conclusions.
- Security data modeling and anomaly detection: model account behavior, access records, and security logs to identify abnormal login, privilege misuse, batch data operations, sensitive-data usage, and policy exceptions.
- Shadow AI and data-fence governance: assess external AI usage risks including sensitive prompt input, file upload, retained conversation context, unauthorized AI service access, and outbound data exposure; connect access control, DLP, audit logs, alerting, and exception handling into executable governance.
- DLP and sensitive-data exposure evaluation: review email outbound flows, file transfers, cross-environment file movement, and sensitive-data handling controls; evaluate product coverage, policy configuration, auditability, and response workflow.
- Security assessment Skill / Agent building: package reusable checklists, risk-decision logic, evidence requirements, output templates, and assessment workflow automation to improve consistency and delivery efficiency.
- Security governance platform support: migrate security logs/data into a centralized governance platform, design SQL/JSON models, support dashboard-based risk visibility, and structure data for downstream investigation and reporting.
- API security and authentication: evaluate API gateway authentication/authorization expectations, token handling, access boundaries, logging, and compliance-oriented control design.

## Security Domains
- Data security governance, DLP, sensitive-data classification, and exposure review
- Identity and access control, PAM/bastion/jump-host audit, least privilege
- Cloud security assessment, logging, monitoring, and evidence collection
- API gateway security, authentication/authorization, and auditability
- Shadow AI, AI usage governance, prompt/file outbound risk control
- Security analytics, anomaly detection, behavior modeling, and control-effectiveness review

## Data / Engineering Stack
- Python: Pandas, scikit-learn, TensorFlow, FastAPI, Pydantic, pytest
- R: ggplot2, dplyr, tidyr
- SQL: analytical modeling, joins, aggregation, audit/risk queries, OLAP workflows
- Data platforms: BigQuery, Looker Studio, Tableau, Power BI, Advanced Excel
- Cloud: AWS S3, EC2, CloudTrail; GCP BigQuery; Azure fundamentals
- AI / automation: RAG, LLM interfaces, tool-calling agents, agentic workflow orchestration, security assessment assistants, and context-aware tool selection
- Engineering basics: Linux, Git, HTTP/API security, IAM, authentication & authorization, access control, security event analysis, cloud security, and secure system assessment

## Applied Analytics
- User behavior analytics, temporal event correlation, segmentation, lookalike modeling
- Risk indicators, anomaly rules, before/after policy-effectiveness analysis
- ETL/ELT workflows, SQL/JSON modeling, OLAP analytics, dashboarding, KPI design, user behavior analysis, and security data analytics

## Projects
- RAG LLM Interface Trial: retrieval-augmented LLM interface for context enhancement and report retrieval.
- Shadow AI Data-Fence PoC: outbound prompt/file-risk redaction concept using FastAPI, Presidio, custom recognizers, and browser-extension UX.
- Fund-System Security Architecture: least-privilege architecture around application, database, bank-connectivity, DMZ, bastion/PAM, logging, and monitoring controls.

## Languages
- Mandarin: Native
- Cantonese: Fluent
- English: Proficient
- Spanish: Beginner

## Good At
- Turning vague security concerns into structured control checklists and evidence requirements
- Bridging security engineering, data analytics, and business-risk communication
- Building lightweight prototypes that make risk visible, testable, and explainable