LLM published work/code to be run on a cluster of GPUs

WARNING: ONLY CONTACT ME IF YOU ARE SERIOUS AND READ THE POST CAREFULLY
-------------------------------------------------------------------------------------------------------------------------
I need to re-run a published paper work, and precisely, the BKL model using different alpha/gamma parameters as shown in Table#3 in the experiment section (see attached table and paper).

The project involves:
- Understanding published paper code in github: https://github.com/mahaucsb/mSPLADE . Basically, - It's a hybrid of two models: Splade (MLM Bert based retrieval model) and Colbert (HuggingFace ranking model) - already done.
- Running the code on a slurm/linux cluster

Ideal candidates should have:
- Experience in understanding complex LLM code - must
- Prior work with the Splade and Colbert models - or any LLM model
- Familiarity with slurm clusters - optional


I will provide access to the cluster with the code/scripts from a former colleague. Please note that all necessary dependencies and libraries are already configured.