quickly discovered that it was extremely difficult to acquire these resources on-demand and at scale in our cloud environments. Further, looking to other third parties presented its own set of risks and challenges.

We turned to听(AWS), which has long been a trusted partner in secure and scalable solutions, to get early access to . With our computing platform acquired, we were ready to roll up our sleeves and do the hard work of exploring how to optimally train and finetune models to our domain. In our first phase of experimentation, we peaked at 16听 compute instances, 128 A100 GPUs, with the longest job taking 36 days to complete training a 70 billion parameter model.

Initial results of our custom models look promising and our research continues, supported by the release of . Our explores the journey that 成人VR视频 took to enable cutting-edge research in training domain-adapted LLMs using听Amazon SageMaker HyperPod.

This is a guest post from John Duprey, distinguished engineer, 成人VR视频.

While it’s easy to be captivated by flashy AI demonstrations, the challenge lies in developing reliable, scalable solutions that deliver tangible value to customers. At 成人VR视频, we’ve embraced this challenge head-on.听听

Effective development of AI solutions at scale requires a breadth of skills, software and infrastructure components with a significant depth of detailed knowledge. Building all this expertise in-house and keeping it up to date is barely possible.听

The 成人VR视频 collaboration with Amazon Web Services (AWS) has been important in allowing us to build a solid, customized toolchain while also providing feedback and proposals on how to optimize AWS offerings to better meet our needs. On the , we shared extensive details on how 成人VR视频 achieved AI/ML innovation at pace with machine learning operations (MLOps) services on the AWS SageMaker ecosystem.听

The generative AI and MLOps spaces are still early stage and fluid. As an engineer within 成人VR视频 Labs 鈥� the dedicated applied research division of 成人VR视频 鈥� I find it exciting to bring stability and solidity to this challenging, fast-paced environment. I hope you enjoy learning about 成人VR视频 Labs鈥� .听

This is a guest post from Danilo Tommasina, distinguished engineer, 成人VR视频.听

Tax research has historically been a labor-intensive process, often requiring extensive manual effort and expertise, combined with the pressure of ensuring accuracy to enable professionals to take and defend tax positions with confidence. The introduction of Checkpoint Edge with CoCounsel gives tax professionals the ability to leverage the transformative capabilities of AI-Assisted Research. Through dialogue with a GenAI assistant, tax professionals can simply ask a question in everyday language. The solution will provide a relevant answer with links to trusted Checkpoint Edge editorial content and source materials in moments.

鈥淩esearch is essential for tax professionals and getting it right can take hours; it is hard and time-consuming,鈥� said Nancy Hawkins, vice president of Product Management, Research, 成人VR视频. 鈥淵et accountants are not lawyers trained in interpreting legislation, and the workforce is shrinking. That is why a GenAI-powered solution has the power to transform the profession. GenAI enables accountants to get it right, fast.鈥�

When it comes to how GenAI can help reduce the burden, 77% of tax professionals think GenAI can be applied to professional work, with increased efficiency and productivity being rated the top reason it should be applied to tax work according to the听2024 Generative AI in Professional Services听report by 成人VR视频.

Trusted technology the tax industry can rely on

Tax firms (83%) and corporate tax professionals (71%) rate technology companies as the primary influencers on the future of GenAI. With influence comes responsibility, and 成人VR视频 works within a rigorous set of听AI and data principles听to ensure trustworthiness in its design, development, and deployment of AI in its solutions.

To expand access to GenAI by unlocking its potential in tax research for its customers, 成人VR视频 built Checkpoint Edge with CoCounsel securely using Amazon Bedrock. Amazon Bedrock is a fully managed service that offers a choice of high-performing foundation models from leading AI companies via a single API, along with a broad set of capabilities for building GenAI applications that meet the high 成人VR视频 standards for security, privacy, and responsible AI.

The use of Amazon Bedrock enabled the 成人VR视频 engineering team to accelerate AI model deployment time from days to hours. The solution is rooted in 成人VR视频-verified databases to deliver reliable, accurate results, while keeping customer data private and secure.

Matt Wood, vice president of Artificial Intelligence Products, AWS said: 鈥淕enerative AI has the potential to revolutionize industries, significantly enhancing operational and cost efficiency. By choosing Amazon Bedrock to build and scale generative AI applications like Checkpoint Edge with CoCounsel, 成人VR视频 can access leading foundation models and customize them using their proprietary data, with the confidence that this data is secure and private by default. We look forward to continuing our work with 成人VR视频 to support generative AI-powered tools that increase productivity and boost innovation.鈥�

Key features of Checkpoint Edge with CoCounsel

Checkpoint Edge with CoCounsel empowers tax professionals with trusted answers in a simple chat interface.

• Trusted answers:听The solution enhances customers鈥� efficiency by delivering synthesized, easily digestible answers along with links to Checkpoint Edge editorial content and source materials, helping professionals spot key risks and resolve questions using powerful summaries from thousands of documents in a matter of seconds.

Since AI-Assisted Research is grounded in 成人VR视频 curated, vetted, and up-to-date Checkpoint Edge content, customers can be assured that they are getting trusted guidance and expert analysis.

• Simple chat interface:听The solution leverages large language models (LLMs) to provide a conversational interface to deliver straightforward answers to customers鈥� questions, as if they were speaking directly to a trusted subject matter expert or advisor. With an easy-to-use, intuitive interface, AI-Assisted Research can be successfully used by tax professionals at all skill levels.

In addition, by synthesizing the most accurate answers from pertinent sources, the solution empowers junior tax professionals to conduct research faster and with confidence, without an over-reliance on the expertise of senior colleagues.

Available for purchase now as part of an early adopter program, Checkpoint Edge with CoCounsel is expected to debut initially in the United States, starting from summer 2024.