Student - Math Research Project Assistant

Ursinus CollegeCollegeville, PA
3d

About The Position

The Math, computer Science, & Statistics departments are seeking an organized and motivated individual to support ongoing research at the intersection of artificial intelligence, mathematics education, and human-centric computing. The project integrates Retrieval-Augmented Generation (RAG), large language models (LLMs), and educational data systems and pedagogical approaches from the MathForum and EnCoMPASS. Students will contribute to building systems that generate and evaluate intelligent feedback for students’ mathematical problem-solving, in collaboration with partners from the 21st Century Partnership for STEM Education and students and faculty from Drexel University. Responsibilities: The project will focus on developing Python-based software for data analytics and prompting AI systems via relevant APIs. Students will contribute to AI system development by implementing RAG pipelines with tools like LangChain, LlamaIndex, HuggingFace embeddings, and Chroma/FAISS, and by running local LLMs through Ollama while comparing them with GPT-4. They will also work on integration tasks such as prototyping backend services for EnCoMPASS and building UI features like an “AI Respond” button. Research and evaluation will include A/B testing of human versus AI feedback, experiment design, data analysis, and documentation for publication.

Requirements

  • Currently enrolled as a full-time student at Ursinus College.
  • Must be dependable.
  • Communicate effectively and protect confidential information.

Nice To Haves

  • Python programming
  • Familiarity with machine learning frameworks and APIs
  • Experience with development tools
  • Comfort with middle- and high-school mathematics

Responsibilities

  • Developing Python-based software for data analytics
  • Prompting AI systems via relevant APIs
  • Implementing RAG pipelines with tools like LangChain, LlamaIndex, HuggingFace embeddings, and Chroma/FAISS
  • Running local LLMs through Ollama while comparing them with GPT-4
  • Prototyping backend services for EnCoMPASS
  • Building UI features like an “AI Respond” button
  • A/B testing of human versus AI feedback
  • Experiment design
  • Data analysis
  • Documentation for publication
© 2024 Teal Labs, Inc
Privacy PolicyTerms of Service