Kaiser Pister
I am a lecturer in the Computer Science
Department at
University of Wisconsin, Madison.
I am CTO and head of AI at All3D.
I am also an AI advisor through
Pister Labs.
As a founder, I have started 3 companies (failed, running, acquired). My work is at the intersection of education, language, and computers,
focusing on applied natural language processing for undergraduate education.
- Extending LLM capabilities to new domains and tasks with In Context Learning
- Automatic management of private AI instances
- Usability of LLM apis. Project: SLLIM
- Efficient retrieval and understanding of large documents
- Programming language development in a post-LLM world. Project LLMC
Teaching
- [Fall 2022 - Spring 2025] - CS 538: Theory and Design of
Programming Languages
This core course forces students to adopt a new perspective on programming. We dive into concepts such as functional programming design, type systems and memory management. - [Fall 2023 + Spring 2024] - CS 638: Topics in Natural
Language
Processing
This seminar course will focus on LLMs in practice, giving students hands on experience developing with the state of the art research techniques.
Information about my lab and my work
If you are interested in working with me on a project, check out the FAQs below.
Student FAQs
-
I am interested in doing research with you. What are the next steps?
First of all, I must note that I am not a research faculty member of UW. That means that any work done with me will not be for credit and cannot be funded through UW. If you are interested in hacking on hard problems and gaining hands on experience, I am accepting students to work with me on a variety of projects. Please stop by office hours or email me directly to discuss. -
Would you write me a letter of recommendation?
Letters of recommendation are a quid pro quo. I cannot write a worthwhile letter for you based only on your in-class participation and grade. Therefore, I have all students who want a letter of recommendation work with me on a micro-project. This project can vary in scope, topic and difficulty. Please stop by office hours or email me directly to discuss. -
What is course XYZ about? What will we cover? Is there a syllabus?
Stop by on syllabus day to learn more.
Startup FAQs
-
I am starting a new company and seeking advice on the state of machine learning.
Congratulations! I am always happy to chat with new founders and provide guidance where I can. Grab some time on my calendly, and we can chat through your ideas. I can operate as a sounding board or critic, depending on your needs. -
I have started a company, acquired funding, and am seeking some help with development.
Calendly is the best way to get a hold of me. Shoot me an intro email with the thesis of your company and I will be ready to talk shop. -
I am in Madison, Wisconsin and want to grab a coffee.
Great! Send me an email. I love coffee.
Papers and Re-Implementations
- PromptDoctor: Automatic Prompt Diagnosis and Repair
- PromptSet: A Programmer's Prompting Dataset, Published LLM4Code '24
- Selective Annotation: A reimplementation to improve in context learning with GPT-3 for a client.
- BERT: A reimplementation and experiment to gain experience with transformers.
- Character Embedded Siamese Networks: Original work done for my thesis to perform named entity recognition on US patent documents for pharmacies.