Attachement 4 - Representative Technology Stack.pdf

PDF 42 KB Posted

Attached to
AUTOMATED SOLUTIONS (INCLUDING ARTIFICIAL INTELLIGENCE) FOR USPTO
Solicitation number
Not on record
Issued by

About this file

Attachment 4 is a document detailing the United States Patent and Trademark Office (USPTO) preferred technology stack for an upcoming artificial intelligence and automated solutions procurement. The technology infrastructure encompasses multiple cloud platforms including AWS (primary), Azure, and Google Cloud, with a software development methodology based on DevSecOps utilizing tools like Rally, Git/GitLab, SonarQube, and Nexus for automated deployment pipelines.

For artificial intelligence capabilities, the USPTO seeks access to advanced AI models across multiple platforms, including Amazon SageMaker, Bedrock, Azure AI, and Google Vertex AI. Specific AI models of interest include Meta's Llama, Anthropic's Claude, OpenAI's GPT, and Google's Gemini. For machine learning projects, the agency will use MLOps with MLFlow, and data management will be handled through Databricks on AWS. This technology stack represents the agency's comprehensive approach to integrating cutting-edge AI and cloud technologies to enhance patent and trademark examination processes.

View the file

Other files

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

Attachment 4: Representative Technology Stack

The USPTO technology stack is broad. For the purposes of this RFI, the following is the preferred tech stack:

Infrastructure: AWS (Primary cloud), Azure and Google Cloud Platforms.

Software development methodology: USPTO DevSecOps with Rally, Automated Deployment pipelines with Git/GitLab, SonarQube, Nexus etc.

Artificial Intelligence: Amazon SageMaker and Bedrock with access to frontier models (including Meta’s Llama, Anthropic’s Claude), Azure AI (including OpenAI’s GPT and o – reasoning models), Google Vertex AI and Gemini. For classical ML projects - MLOps with MLFlow.

Data: Databricks (on AWS)

File details come from the government source that posted it. Updated .