Research Assistant (LLM and Reinforcement Learning)

Research Assistant (LLM and Reinforcement Learning) Position Overview The Hong Kong University of Science and Technology (Guangzhou) invites applications for Research Assistant positions in LLM and Reinforcement Learning. The project will develop an artificial intelligence agent that learns how to select effective and responsible interventions for consumer-loan delinquency problems. We are looking for enthusiastic RAs to work closely with the supervisor Prof. Kelvin Yuen and/or his collaborators. Prof. Yuen is an assistant professor at HKUST(GZ), and an assistant professor of economics at HKUST. The successful candidates will work at the intersection of reinforcement learning, large language models, causal inference, and economic modeling. Key responsibilities

1. 01

   Build secure and reproducible pipelines for processing conversation transcripts

2. 02

   Develop representations of borrower states, including latent-state or belief-state representations appropriate for POMDPs.

3. 03

   Implement and train reinforcement-learning models using historical, primarily offline, interaction data.

4. 04

   Evaluate learned policies through off-policy evaluation, simulation, robustness tests, and comparison with human collector decisions.

5. 05

   Conduct model diagnostics, ablation studies, sensitivity analyses, and reproducibility checks.

6. 06

   Work with researchers conducting heterogeneous treatment-effect estimation and causal analysis.

7. 07

   Other ad-hoc tasks and logistics duties assigned

8. 08

   Need to work onsite in Guangzhou and Shenzhen

Qualifications

1. 01

   A bachelor’s or master’s degree in computer science, artificial intelligence, data science, statistics, operations research, or a closely related discipline.

2. 02

   Strong programming skills in Python.

3. 03

   Strong written and verbal communication skills in English and Chinese.

4. 04

   Hands-on experience implementing and training reinforcement-learning models.

5. 05

   Knowledge of Markov decision processes and partially observable Markov decision processes, including state or belief representation, policy learning, reward design, and sequential evaluation.

6. 06

   Experience with a major machine-learning framework.

7. 07

   A solid understanding of probability, statistics, optimization, and machine learning.

8. 08

   Experience working with large, complex, or sequential datasets.

9. 09

   Strong analytical and problem-solving skills. Benefits

10. 01

    Salary is highly competitive and commensurate with experience and education.

11. 02

    Involvement in research projects aimed for top-ranking journals.

12. 03

    RAs demonstrating exceptional performance could receive coauthorship on the project(s). Application Procedure Interested candidates should submit

13. 01

    CV/resume;

14. 02

    brief cover letter detailing relevant experience and interest in the position; and

15. 03

    availability. Application will be open until the position is filled. Please send your applications to kelvinyuen@ust.hk, and we will contact shortlisted candidates within two weeks upon receiving the application. If you have any questions, feel free to contact us anytime.