LLM / GenAI Engineering Leader in Financial Services
I lead enterprise GenAI, NLP, and real-time decision system development at American Express, with a focus on turning advanced machine learning research into production AI capabilities for commercial marketing, sales, and customer engagement.
My work sits at the intersection of large language models, retrieval, recommendation systems, real-time personalization, and enterprise AI platform design. I build and lead systems that help business teams make faster, more relevant, and more intelligent decisions in regulated financial-services environments.
Current Focus
- Enterprise GenAI systems: LLM-powered assistants, sales intelligence workflows, prompt/RAG systems, and applied agentic AI capabilities.
- Real-time decisioning: AI systems for arbitration, targeting, personalization, and customer engagement across digital channels.
- Commercial marketing intelligence: Models and platforms that connect customer intent, product relevance, and outreach strategy.
- Research-to-production AI: Translating ideas from NLP, reinforcement learning, optimization, and statistical learning into scalable deployed systems.
Selected Impact
- Led GenAI and real-time marketing R&D initiatives for Global Commercial Services at American Express.
- Built LLM-powered sales and campaign optimization capabilities using modern foundation models and retrieval-augmented generation.
- Developed real-time AI decision systems on cloud infrastructure, including production capabilities for marketing and sales channel optimization.
- Advanced NLP and intent modeling systems using BERT, XLNet, RoBERTa, Llama, GPT, and related transformer architectures.
- Received the Edward P. Gilligan Award for Innovation, American Express’s highest enterprise-level innovation honor.
Technical Depth
My technical background spans large language models, natural language processing, reinforcement learning, recommendation systems, time-series modeling, statistical machine learning, optimization, and cloud-based AI deployment.
I work primarily with Python, PyTorch, Hugging Face Transformers, OpenAI APIs, pandas, cloud infrastructure, and production ML system design. I also mentor data scientists and collaborate across product, engineering, marketing, and business teams to move AI systems from research and experimentation into reliable enterprise use.
Research Foundation
Before moving into enterprise AI leadership, I earned my Ph.D. in Electrical Engineering and Computer Science from the University of Michigan, where my research focused on optimization, statistical estimation, and intelligent decision systems under uncertainty.
My research has been published in venues including ICLR, IEEE Transactions on Automation Science and Engineering, the American Control Conference, and the International Journal of Production Research.
Selected publications:
K. Liu. “SetCSE: Set Operations using Contrastive Learning of Sentence Embeddings.” International Conference on Learning Representations (ICLR), 2024. [OpenReview]
K. Liu, N. Li, I. Kolmanovsky, A. Girard. “A Vehicle Routing Problem with Dynamic Demands and Restricted Failures Solved Using Stochastic Predictive Control.” American Control Conference, 2019. [IEEE]
P. Alavian, Y. Eun, K. Liu, S. M. Meerkov, L. Zhang. “The (α, β)-Precise Estimates of MTBF and MTTR: Definition, Calculation, and Observation Time.” IEEE Transactions on Automation Science and Engineering, 2020. [IEEE]