Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm a PhD candidate in Machine Learning focused on interpretable AI and scalable data analytics. I design modular multi-agent AI systems with chain-of-thought reasoning to deliver reliable, transparent solutions for enterprise video and data understanding. I bridge academia and industry, building end-to-end data platforms with Databricks, PySpark, and cloud infrastructure while staying grounded in risk management, regulatory compliance, and stakeholder alignment.…I'm a PhD candidate in Machine Learning focused on interpretable AI and scalable data analytics. I design modular multi-agent AI systems with chain-of-thought reasoning to deliver reliable, transparent solutions for enterprise video and data understanding. I bridge academia and industry, building end-to-end data platforms with Databricks, PySpark, and cloud infrastructure while staying grounded in risk management, regulatory compliance, and stakeholder alignment.

Patrick Duan

Data Scientist, AI Engineer, AI Strategy Consultant, +5





I'm a PhD candidate in Machine Learning focused on interpretable AI and scalable data analytics. I design modular multi-agent AI systems with chain-of-thought reasoning to deliver reliable, transparent solutions for enterprise video and data understanding. I bridge academia and industry, building end-to-end data platforms with Databricks, PySpark, and cloud infrastructure while staying grounded in risk management, regulatory compliance, and stakeholder alignment.…I'm a PhD candidate in Machine Learning focused on interpretable AI and scalable data analytics. I design modular multi-agent AI systems with chain-of-thought reasoning to deliver reliable, transparent solutions for enterprise video and data understanding. I bridge academia and industry, building end-to-end data platforms with Databricks, PySpark, and cloud infrastructure while staying grounded in risk management, regulatory compliance, and stakeholder alignment.

Available to hire

I’m a PhD candidate in Machine Learning focused on interpretable AI and scalable data analytics. I design modular multi-agent AI systems with chain-of-thought reasoning to deliver reliable, transparent solutions for enterprise video and data understanding. I bridge academia and industry, building end-to-end data platforms with Databricks, PySpark, and cloud infrastructure while staying grounded in risk management, regulatory compliance, and stakeholder alignment.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Language

English

Fluent

Work Experience

Risk Data Analyst

April 1, 2025 - November 11, 2025

Automated large-scale data workflows on Databricks for efficient processing and analysis of trade and account activity. Built Python and SQL-based anomaly detection pipelines to surface irregular patterns and automate EA activity detection in system. Developed interactive dashboards to visualize key operational metrics and support data-driven decision-making. Collaborated on rapid prototyping of predictive analytics solutions for intelligent alerting and process automation.

Machine learning Intern at TikTok

September 1, 2025 - September 1, 2025

Developed scalable, interpretable multimodal AI frameworks for automated content moderation and abuse detection, leveraging large language models (LLMs), visual transformers (ViT), and graph neural networks (GNN) to analyze complex social media data and behavioral signals. Designed hybrid reasoning pipelines that combine semantic search, rule-based filtering, and retrieval-augmented generation (RAG) to enhance triage and detection of nuanced, high-risk content in both short- and long-form videos. Prototyped graph-enhanced safety systems by integrating user interaction graphs and temporal event modeling with foundation models, enabling early identification of coordinated misinformation, forgery, and emerging security threats. Explored fairness, adversarial robustness, and explainability in classification systems for trust & safety, bridging advances in prompt engineer

Tutor at University of New South Wales

May 1, 2025 - May 1, 2025

Delivered practical workshops and tutorials on advanced analytics, AI techniques, data visualisation, and cloud-based data engineering solutions. Mentored students in designing, training, and deploying machine learning models, focusing on real-world data applications and best industry practices.

Tech Consultant at Ernst & Young

April 1, 2023 - April 1, 2023

Led advanced data analysis projects to assess technology risks, performing extensive data transformation, validation, and migration tasks on cloud platforms (AWS/GCP). Designed and optimised complex SQL and PySpark queries for performance tuning, significantly improving audit accuracy and data-driven decision-making. Developed automated risk prediction models, incorporating predictive analytics and anomaly detection methods, enhancing clients' operational resilience and regulatory compliance. Facilitated strategic workshops, clearly presenting data-driven insights to stakeholders, influencing critical business decisions.

Risk Data Analyst

April 1, 2025 - November 12, 2025

Automated large-scale data workflows on Databricks for processing and analysis of trade and account activity; built Python and SQL-based anomaly detection pipelines; developed interactive dashboards; supported rapid prototyping of predictive analytics for intelligent alerting and automation.

Researcher / PhD Candidate - Modular Multi-Agent Reasoning for Long-Form Video Understanding at University of Technology Sydney

January 1, 2025 - November 12, 2025

Developed a modular multi-agent AI framework for long-form VideoQA, integrating retriever, planner, and verifier agents with chain-of-thought reasoning; designed interpretable collaboration protocols and graph-based event segmentation to support compositional reasoning and transparent decision-making.

Researcher - Dynamic Risk Prediction in Financial Networks (AI & Analytics) at University of New South Wales

May 1, 2025 - May 1, 2025

Developed a time-aware deep reinforcement learning framework with Graph Neural Networks (DySAT, TGCN) for real-time financial risk insights; built and managed complex PySpark data pipelines on Databricks; created interactive visualizations to improve stakeholder interpretability.

Researcher - Continual Learning in Generative Systems at University of Technology Sydney

January 1, 2025 - November 12, 2025

Developed scalable LLM systems with persistent memory, retrieval-augmented generation, LoRA tuning, and modular forgetting mechanisms; designed curriculum-based pretraining and memory management strategies for continual learning.