1.1      Initiative Summary:

Businesses looking to operationalize LLM-supported applications will benefit from using cloud-based (private or public) LLM “as a service” (LLMaaS) platforms for governance and scalability. Among many features, data governance (primarily for unstructured text) will be a critical offering of these platforms, including that from Blattner Technologies. This initiative will focus on contributing to the development of an extensible end-to-end data governance framework, including external data ingestion, parallelized data preparation and analytics, and versioning.

 

1.2      Desired Outcomes

-   Prototype innovative workflow-based capabilities for preparing unstructured text in a scalable, traceable, and intuitive manner for downstream LLM-related tasks, such as training and fine-tuning.

-   Presentation to broader company highlighting approach, challenges, solutions, and significant insights stemming from the effort.

 

1.3      Core Skills Required

-   Required skills:

o   Fundamental LLM knowledge (e.g., prompt engineering, fine-tuning)

o   NLP-based development (e.g., tokenization, embedding generation, and operations, textfication)

o   Python development

o   Experience with parallel distributed systems and/or parallel computation libraries such as Spark, Dask, or RAPIDS

-   Optional/preferable skills:

o   Kubeflow

o   Vector databases

o   Experience with NLP libraries such as spaCy and gensim

 

1.4      Estimated Effort

-   Full-time summer internship (40 hours/week)

-   Depending on progress, work may extend to part-time during the Fall semester (e.g., 10 hours/week)

 

1.5      Additional Information

This is a remote internship opportunity, working with summer mentors and reporting to the Chief Product Officer of BOSS AI. The group has a deep focus on implementing LLMs “as a service” (LLMaaS) and team members have a range of skills from enterprise software engineering, NLP, ML, and UX. You can expect to gain valuable experience in operationalizing LLMs and addressing critical security needs for all language models.