Liu Liu

When Assistant Professor of Marketing Liu Liu began teaching in the spring of 2022, she didn鈥檛 anticipate having to update the curriculum almost immediately. But with the whirlwind of advances in Generative AI, large language models and deep learning (see 鈥淭he Vocabulary of AI鈥), course adaptations are a regular requirement.
Despite those 鈥渂ig technological jumps,鈥 as Liu describes them, her love of teaching and mentoring remains constant. She finds the ability to make an impact on students鈥 lives at a critical stage of their growth鈥攚hen they are curious and exploring opportunities and possibilities鈥攊ncredibly rewarding.
鈥淚 am also very passionate about sharing the knowledge that I have in the most intuitive way,鈥 she said.
Liu also values the continuous learning and intellectual freedom afforded by academic research. It allows her to 鈥渃hoose what topics to pursue and really dive deep into them.鈥
Connecting AI to business strategy
Liu was the first faculty member at Leeds to have a course fully dedicated to bridging AI and business鈥攁 program so impactful that even non-Leeds students at CU reach out expressing interest in enrolling. Today, Leeds offers more than 45听. Liu also teaches an undergraduate marketing course on pricing and channels of distribution.
Her teaching philosophy extends beyond theory.
鈥淚 don鈥檛 just teach the theories of deep learning鈥攍ike this model or that model. I also do case studies with students where we discuss research papers that really apply AI to solve a business problem.鈥

鈥淚 don鈥檛 just teach the theories of deep learning ... I also do case studies with students where we discuss research papers that really apply AI to solve a business problem.鈥
Liu Liu, Assistant Professor of Marketing
Through her 鈥渆nd-to-end pipeline鈥 approach, students in her AI course tackle every stage of problem-solving鈥攆rom defining business challenges to selecting models, interpreting results and exploring implications. Students even engage in hands-on coding exercises to replicate results, reinforcing both technical and strategic understanding.
鈥淚n my opinion, this is the strength of business analytics compared to pure data science or a computer science program.鈥 By applying the data to specific domains, such as behavior marketing, students connect the right questions to the right problem and then find the right tool for the solution, she said. 鈥淭hat is my goal鈥攖o prepare students for that.鈥
From Google to the classroom

After completing her undergraduate and master鈥檚 degrees in computer science, Liu headed to Google Pittsburgh, where she worked for three years as a software engineer.
鈥淚 feel very fortunate that was my first job,鈥 she said. Her work predicted consumer ad click-through rates for what is now called the Google Display Network. She applies that experience all the time in the classroom, she said, as she teaches about using large-scale machine learning in relation to consumer behavior.
In the research realm, Liu specializes in empirical marketing. Her research straddles the intersection of marketing and machine learning in such areas as visual marketing, branding, consumer preference measurement, product design and social media.听These insights feed into her teaching, connecting the worlds of computer science and business in an interdisciplinary way.
鈥淚 am just fascinated by consumer preferences and consumer behavior,鈥 she said, explaining that the technology allows companies to measure consumer preferences to predict their choices and accommodate them with better designed products and branding.
Keeping pace with AI
鈥淎I is a high-speed train that will not stop, so we better think about how to get on the train and figure out how to drive it,鈥 said Liu.
鈥淢y perspective has always been that you have to be able to use AI correctly and responsibly.鈥 She shared a metaphor: 鈥淚f you don鈥檛 know how to operate, don鈥檛 be a doctor.鈥

鈥淎I is a high-speed train that will not stop, so we better think about how to get on the train and figure out how to drive it."
Liu Liu, Assistant Professor of Marketing
To use AI effectively, her courses emphasize the correct use of the technology. 鈥淵ou have to be able to understand the method, the tool and the model to apply the right numbers, do the math correctly, and supply the right implications,鈥 she said. That鈥檚 part one, she explained.
Part two relates to the ethical use of AI and the idea that technology does not exist in isolation. Technology is always associated with human decision-making and its consequences, she said.
A journey from China to Colorado
Originally from Tianjin, a city near Beijing, Liu has been in the United States since 2008, what she considers a defining stage in her life during which she truly became an adult (earning her own money and paying taxes, she laughed).
That life experience helps Liu relate to her students as they navigate similar transitions into adulthood and professional careers.
Outside the classroom, Liu added another milestone to her adult life鈥攐ne that requires just as much energy as any AI model and comes with its own form of deep learning鈥攔aising twins, a son and a daughter.
Looking ahead
What鈥檚 next? Keeping up with her research and teaching continue to excite Liu as she remains committed to helping students thrive and push boundaries in an AI-driven world.
Figuring out how to balance that with her family life is part of the quest and another exciting challenge entirely. She joked, 鈥淭hat may be an unsolvable problem.鈥
The Vocabulary of AI
Artificial Intelligence
Traditional AI uses algorithms and data to enable computers to perform tasks that mimic human intelligence. It can be narrow (specific tasks) or general (broad capabilities).
Generative AI
GenAI is a subset of AI that creates original content like text, images, music or videos based on natural language prompts.
Machine Learning
A subfield of artificial intelligence that allows computers to learn from data by identifying patterns, requiring minimal human intervention.
Deep Learning
A more mathematically complex evolution of machine learning that teaches computers to learn by example using artificial neural networks modeled on the human brain.
Large Language Models
Large language models (or LLMs) process vast amounts of data to mimic the way humans communicate.
Sources
- University of Colorado,听
- Microsoft,
- MIT Management,听
- Stanford University,听
- USC,