Postdoctoral Fellow - Data Analytics

University of Arkansas • Little Rock, AR

1d

About The Position

The University of Arkansas at Little Rock is a metropolitan research university that provides an accessible, quality education through flexible learning and unparalleled internship opportunities. At UA Little Rock, we prepare our more than 8,900 students to be innovators and responsible leaders in their fields. Committed to its metropolitan research university mission, UA Little Rock is a driving force in Little Rock's thriving cultural community and a major component of the city and state's growing profile as a regional leader in research, technology transfer, economic development, and job creation. The Data Life Cycle and Curation Theme aims to achieve automated heterogeneous data curation. While we have already developed the Data Washing Machine System for Data Quality, our research goal is to create fully automated processes for all phases of the data life cycle. We seek a Postdoctoral Fellow with expertise in Large Language Models (LLMs) to achieve this. This individual will be critical in developing customized LLMs and integrating them with our existing system. Additionally, the expertise of the postdoctoral staff will benefit not only the DART project but also other faculty members who require LLM-related skills. Ultimately, this position will contribute to Arkansas’ economic development by strengthening our capabilities in generative AI. This position is governed by state and federal laws and agency/institution policy.

Requirements

• A Ph.D. in Computer Science, Information Science, or related technology field
• Proof records of success in research, including peer-reviewed publications in Large
• Language Model, Generative Al, and Data Quality
• Experience in customizing open-source large language models, including Llama 2 from Meta and Gemma from Google for data quality
• Good communication and teamwork skills
• Strong skill in large language model customization techniques including prompt engineering, model fine-tuning, and model alignment

Responsibilities

• Create a method for customizing LLMs specifically tailored to data phases such as data blocking, entity matching, and entity resolution within the data lifecycle
• Implement the developed method using open-source LLMs, including tools like Llama from Meta
• Evaluate the effectiveness of the LLM-based solutions using both synthetic and real-world data
• Integrate the LLM solutions seamlessly into the Data Washing Machine for automated heterogeneous data curation
• Develop HPC solutions utilizing high-performance GPU clusters to enhance efficiency and scalability

What This Job Offers

• Job Type: Full-time
• Education Level: Ph.D. or professional degree