OCI Data Science 2024: A year of innovation in AI

2024 has been an incredible year for Oracle Cloud Infrastructure (OCI) Data Science, marked by innovative advancements, powerful new features, and inspiring use cases. Oracle has continued to empower data scientists and developers with tools that help simplify complex workflows, enable cutting-edge AI applications, and deliver actionable insights.

From the launch of AI Quick Actions to significant updates like support for the latest Llama models and integration with Hugging Face, this year has been all about accelerating innovation and making AI accessible to everyone.

2024: A year in review

January: Making generative AI accessible

OCI Data Science kicked off the year with a focus on democratizing AI. The initial introduction of the AI Quick Actions feature enabled users to work with large language models (LLMs) in a code-free environment, making AI more accessible for non-experts. For more information, see Generative AI Made Easy in OCI Data Science.

April: Introducing AI Quick Actions

In April, Oracle officially released AI Quick Actions in OCI Data Science, revolutionizing how foundation models are deployed, fine-tuned, and evaluated. This feature provided a streamlined, code-free environment for working with LLMs, catering to users across all skill levels.

June: Embracing diversity in AI models

The blog post, Deploying ELYZA with vLLM, highlighted deploying ELYZA, a Japanese LLM, in OCI Data Science, showcasing global AI diversity.
OCI Data Science introduced the Bring Your Own Model (BYOM) feature for AI Quick Actions, allowing users to integrate custom models seamlessly with Hugging Face.

July: Advancing LLM deployment

Oracle demonstrated deploying open-source LLM Llama 3.1 405B using the Bring Your Own Container (BYOC) approach, showing flexibility and scalability in working with massive LLMs. For more information, see this blog on Deploying Llama 3.1 405B.

August: Unlocking AI potential with NVIDIA NIM

In collaboration with NVIDIA, Oracle introduced NVIDIA NIM microservices on OCI, offering groundbreaking tools for advanced AI workloads. This work highlights how OCI Data Science helps enable scalable, high-performance machine learning and inferencing workflows using NVIDIA's state-of-the-art technology. For more information, see the blog on Unlocking AI Potential: NVIDIA NIM on OCI.

September: Expanding AI ecosystem

OCI Data Science became one of the first platforms to support Llama 3.2, enabling fine-tuning and deployment for advanced AI projects.

Hugging Face models joined the ecosystem, giving customers access to state-of-the-art LLMs and further expanding AI Quick Actions' capabilities. For more information, see Simplifying AI Integration with Hugging Face.

October: Driving cost-efficient AI

Oracle partnered with Ampere to optimize LLM inferencing on cost-effective Arm CPUs, enabling businesses to deploy affordable AI solutions.

November: Exploring agentic AI

Oracle explored the concept of agentic AI, positioning it as the future of adaptive and autonomous systems, using OCI Data Science as the foundation for these innovations.

December: Introduction of Llama3.3

Bring the newly announced Llama 3.3 70B Instruct model from either Hugging Face or Meta to OCI Data Science easily thanks to AI Quick Actions feature.

Transformative use cases with OCI Data Science

OCI Data Science has proven instrumental in solving real-world challenges across industries, driving impactful outcomes, including the following examples:

• Customer personalization: Build recommendation systems that help enhance user experience in industries like retail and media, delivering tailored content and product suggestions.
• Fraud detection: Deploy AI models to help identify anomalies and suspicious behavior in real time for banking and insurance, which may result in stronger security and a reduction of financial losses.
• Operational efficiency: Automate predictive maintenance in manufacturing, reduce downtime, and optimize supply chains with advanced forecasting techniques.

These examples demonstrate how OCI Data Science empowers businesses to use AI effectively, addressing their unique needs with scalable, cost-efficient solutions. If you're seeking a data science platform that offers flexibility, scalability, and comprehensive support for the entire machine learning (ML) lifecycle, OCI Data Science is designed for you. It includes the following key features:

• Flexible data access: Seamlessly connect to various data sources, including Oracle Autonomous Database, Object Storage, and external databases, helping ensure that your data is readily available for analysis.
• Data preparation at scale: Utilize integrated tools like Apache Spark to process and prepare large datasets efficiently, enabling robust data pipelines.
• Open source machine learning frameworks: Work with popular frameworks, such as TensorFlow and PyTorch, providing the versatility to use tools you’re familiar with.
• Accelerated model training: Utilize powerful GPUs to expedite deep learning model training, achieving performance speedups of 5–10 times faster than CPUs.
• Autoscaling and burstable instances: Automatically scale resources up and down based on workload demands, with support for burstable instances to handle peak loads cost effectively.
• Anomaly detection with time series: Detect anomalies in data using time series constructive anomaly detection methods.
• Collaborative environment: Engage in a JupyterLab based workspace that fosters teamwork among data scientists, which can be applied to enhance productivity and knowledge sharing.
• Comprehensive model management: Utilize the model catalog to preserve, share, and deploy ML models seamlessly, helping ensure reproducibility and governance.
• Integrated model deployment and monitoring: Deploy models efficiently with integrated tools and monitor for model prediction drift and feature distribution changes, helping ensure consistency and reliability over time.
• MLOps integration: Automate and streamline your machine learning workflows with features like managed model deployment, ML pipelines, and continuous monitoring for your model predictions drift.

By incorporating these features, OCI Data Science helps empower you to build, train, deploy, and manage machine learning models effectively, addressing complex data challenges with confidence. For details, see our Data Science GitHub repository.

2024 Key announcements and highlights

We announced the following big items this year:

1. AI Quick Actions: A no-code solution empowering users to deploy, fine-tune, and evaluate LLMs efficiently. See demo.
2. Integration with Hugging Face: Access to popular models like Mistral and Llama, streamlining AI adoption.
3. BYOM and BYOC support: Flexibility for users to bring their custom models and containers.
4. Support for Llama 3.3 and other LLMs: Advancing deployment capabilities with Meta’s state-of-the-art models.
5. NVIDIA NIM on OCI: Unlocking new possibilities for advanced AI workloads with NVIDIA NIM on OCI.
6. Cost-efficient computing with Ampere A1: Enabling budget-friendly inferencing on Arm CPUs.
7. Agentic AI framework: Shaping the future of AI with OCI Data Science at its core.
8. Support for NVIDIA A100, H100, and L40S GPUs across all Data Science resources.
9. Launch of ML Insights of monitoring ML models and data.

Conclusion

2024 has been a transformative year for OCI Data Science, solidifying its position as a leader in AI and ML. From simplifying workflows with AI Quick Actions to enabling groundbreaking use cases with powerful LLMs, Oracle continues to empower data scientists and organizations to innovate.

Whether you’re building personalized customer experiences, detecting fraud, or optimizing operations, OCI Data Science provides the tools and support to solve your toughest challenges with AI.

Start your journey today and discover how Oracle Cloud Infrastructure Data Science can help you unlock the full potential of your data. To learn more, see the following resources:

Try Oracle Cloud Free Trial! A 30-day trial with US$300 in free credits gives you access to Oracle Cloud Infrastructure Data Science service.

For more information, see the following resources:

• OCI Data Science sample repository on GitHub.
• Oracle Data Science service technical documentation.
• Using the Oracle Resource Manager to Configure Your Tenancy for Data Science.
• Star and clone our new GitHub repo! We included notebook tutorials and code samples.
• Watch our tutorials on YouTube.
• Search for Oracle Data Science and try LiveLabs.