Gen AI RFI Questions.docx
DOCX document 26 KB Posted
- Attached to
- GenAI Request for Information-Results Update
- Solicitation number
- 2031JW25N0007
About this file
This is a Request for Information (RFI) question set from the Office of the Comptroller of the Currency (OCC) seeking industry feedback on their Generative AI Tool Support Services requirement. The document contains 16 questions across four categories: pricing, technical tasks, PWS feedback, and general PWS considerations.
The questions focus on gathering information about contract structuring (fixed-price vs. labor-hour), typical staffing levels for AI tool operations and maintenance, technical aspects of GenAI development including model training, data processing, user interface design, and security practices. The questions also seek input on performance metrics, task organization, and the completeness of the draft PWS. Respondents are required to provide detailed responses beyond simple yes/no answers, with submissions due by December 19, 2024 at 5:00 PM Eastern to Casey.Helman@occ.treas.gov. This document is part of a larger RFI package that includes a draft PWS which is referenced but not included in this file.
View the file
Other files
| File | Type | Posted |
|---|---|---|
| GenAI-draft PWS_RFI-clean.docx | DOCX document | |
| GenAI-draft PWS_RFI.docx | DOCX document |
On GovTribe
Work with this file
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
Gen AI RFI Questions Pricing Questions
1. How feasible is it for the O&M segment of Generative AI tools to be structured as a Fixed-Price contract?
2. What level of effort (hours per week, Full-Time Equivalents, hours per month) typically characterizes the O&M of such tools?
3. If LH is more appropriate, what labor categories and experience levels would fit this work?
Task Questions
4. What are some of the common activities and tasks involved in the Operations and Maintenance (O&M) of custom Generative AI tools developed in-house?
5. What challenges and considerations arise when inheriting a prototype developed by a different team for full production development? Outline typical activities and tasks for advancing such a prototype to a production-ready state, and what differences exist between cloud-based and on-premises hosting environments.
6. What technologies and frameworks are commonly leveraged in the development of Generative AI solutions, and how is compatibility and seamless integration with existing infrastructure achieved?
7. Which methodologies does the industry typically employ for training and fine-tuning custom Generative AI models? How are ongoing updates and maintenance managed to ensure model accuracy and performance?
8. What are some of the industry-best practices for data ingestion and pre-processing when working with Generative AI tools?
9. What user interface options are commonly adopted in the industry to enhance the usability of Generative AI solutions?
10. What best practices does the industry follow to secure and protect the privacy of data used in Generative AI services?
11. How do proprietary Generative AI solutions scale to handle increasing data volumes and a growing user base?
General PWS Questions
12. Describe the various innovative performance metrics that industry leaders are using to evaluate the accuracy and overall effectiveness of their tools.
13. Considering the draft PWS includes three task areas, are there additional areas or tasks that should be included to comprehensively cover the development and maintenance of AI/ML tools?
14. Conversely, do any of the tasks make more sense as a separate contract, if so which ones and why?
15. Are there other tasks or considerations that should be included in the PWS, if so what are they and why?
16. Is there enough information in the draft PWS to understand the work needed and provide a price model? If not, what additional context would be helpful?
File details come from the government source that posted it. Updated .