Overview

Transforming the immense potential of generative AI and data is crucial yet challenging for startup, SMB, and enterprise initiatives. That’s why AWS and Successive have created the GenAI PoC, where you can get expert guidance and direction on using Large Language Models (LLMs) on AWS. Our AWS Generative AI Proof of Concept Offering is designed to quickly prototype Generative AI use cases to validate the use-case against viability, ROI, technical feasibility, and LLM model functionality and a way to showcase the value of Generative AI for your business.

How We Work with You:

• Define PoC requirements, work with you to establish KPIs, success criteria
• Work with your team to validate GenerativeAI use-case and business requirements
• Analyze the data availability, data sources, and quality, model training infrastructure readiness model training, and deployment
• Set up PoC infrastructure and environment for model training and deployment
• Identify and leverage relevant foundation models based on the use cases and data availability, e.g., GPT 3.5/4.0, Llama 2, etc.
• Leverage vector database to load your business-specific data to provide in-context prompting (optional)
• Model fine-tuning, training on custom data, and prompt optimization to specific use case
• Up to 8 weeks of PoC in phases: Readiness (1 week) and PoC execution (7 weeks)
• This PoC will cover up to one Generative AI use case
• We will work with AWS to get the necessary funding to offset the PoC services cost

Deliverables:

• Working PoC and Demo
• Documented problem statement, Generative AI use case, and KPI and success criteria
• High-level PoC design and technical documentation
• All the software code, prompts, and fine-tuning developed for the PoC
• Documented roadmap to scale and move the PoC to MVP

Outcomes:

• Working prototype for the use case to demonstrate the value to business
• MVP roadmap and execution strategy for the next phase
• ROI and business use case value justification