Alfred Ekoka

Don't shy away from fine-tuning ! Don't believe those who tell you it is super hard and super expensive. It is not. And it can bring in great value once your chain is in prod and you have a matured train/eval dataset ! An amazing article from my friends Narjes Majdoub, Ellen Hirt (Hoeven) and Giran Moodley 📚 What You'll Learn: - When fine-tuning is the right approach. - Step-by-step instructions for preparing your data and environment. - Best practices for running and evaluating fine-tuning experiments. - How to deploy and monitor your fine-tuned models. hashtag #Databricks hashtag #FineTuning

13

2 commentaires

AI Agents are imperative Generative AI based utilities. Multi AI Agents are enterprise class requirements - include functions, code interpreters, data, instructions and models in a declarative fashion. Multi-modal Multi-agentic Copilot framework is what I have created tonight, with a practical hands-on example. Have fun ! #generativeai #artificialintelligence #aiagents #copilot #azure

15

The Architecture of Apache Druid #apachedruid #olap #history #architecture #storage #distributedsystems https://lnkd.in/egqai_ui

1

Substring Extraction in COBOL - Extract a Substring from Variable length Position (Session-38) : In this video, we demonstrate how to extract substrings based on variable length in COBOL programs, helping you manage string manipulation efficiently. Whether you are working on legacy systems or modern applications, this tutorial will guide you through step-by-step techniques for handling substring extraction using COBOL's powerful string functions. Topics Covered: 1.Substring Extraction in COBOL: Learn the fundamentals of how substring extraction works in COBOL and its importance. 2.Handling Variable Length Substrings in COBOL: Understand how to deal with variable-length data fields for more dynamic and flexible COBOL programs. 3.COBOL String Functions Explained: Get an overview of key COBOL string functions like INSPECT and REFERENCE MODIFICATION to enhance your substring extraction tasks. 4.Practical COBOL Example for Substring Extraction: Walkthrough of a real-world COBOL code example to extract substrings based on variable length. Session in English : https://lnkd.in/gnXNnvDs Session in Hindi : https://lnkd.in/gBU_ra3V Tech Sharmit Digital Course Book [Free for everyone] : https://lnkd.in/dsqcaWpA 🔔 Subscribe to Tech Sharmit for more COBOL programming tutorials, tips, and tricks. Stay ahead with the latest insights in COBOL and other programming languages! 👍 Like, Comment, and Share this video to help others improve their COBOL skills and advance their programming knowledge! #TechSharmit #Sharmit #IBMChampionAmitSharma #IBMChampion #ibmi #ibm #Advocacy #cobol #ibmi #as400 #iseries #CobolProgramming #CobolLanguage #CobolCommunity #IBMRisingChampions #IBMTechXchange #hkhhm

2

Remember back in 2012, when PostgreSQL introduced JSON support ? https://lnkd.in/e_Qb8rpc Now in PostgreSQL 17 we have JSON_TABLE and new JSON functions !

3

I wrote a new blog post that answer some of the questions around embedding models, e.g. what are different embedding models? How are they pre-trained? What is MTEB leaderboard? How are these models benchmarked? How to choose a baseline embedding model for your RAG and where to go from there. Check it out 👇

38

1 commentaire

What's AIBI genie ? How does it work ? Who should use it ? Chao Cai Sr Director of Engineering at Databricks answers all this questions and share some best practices as well. If you have not tried it yet after this session you will definitely do it. If you're a Data Analyst, you shouldn't consider genie as threat it's more a companion in the data democratization journey. #Databricks

20

Missed the Snowflake Summit 2024 in San Francisco ? Don't worry ! Join us on July 17th for our webinar : The Best of Snowflake Summit. In this session, you'll: ✨ Discover new features and capabilities announced at Snowflake Summit 2024 🤖 Learn about the latest in AI, genAI, Apache Iceberg, and more 💡 Gain insights into transforming your organization with data and AI Don't miss this opportunity to stay ahead with the latest innovations. Register now! 👉 https://okt.to/qOc9iX #AIDataCloud #SnowflakeSummit #AI #Snowflake

7

1 commentaire

Performance-wise, the new protocol has reduced dataplane latency by 11% on average, and by 17.4% at the P90 mark. In addition to making SQS itself more performant, this change benefits services that build on SQS as well. For example, messages sent through Amazon Simple Notification Service (Amazon SNS) now spend 10% less time “inside” before being delivered. Finally, due to the protocol change, the existing fleet of SQS hosts (a mix of X86 and Graviton-powered instances) can now handle 17.8% more requests than before. #AWS #AmazonWebServices #AWSBlogs #Cloud #CloudComputing #Serverless #SQS

2

[ 🎉 Announcing the Forward Data Conference: Shaping the Future of Data and AI in France and Europe] Over the past three years, Modern Data Network has been dedicated to fostering synergies within the data ecosystem in France. We have experimented with various event formats, ranging from private webinars to coaching sessions and select public meetups, maintaining an exclusive atmosphere for our members. This year, we are excited to broaden our contribution to the ecosystem with a conference designed to address the current challenges faced by data practitioners. The Forward Data Conference promises to be a must-attend event, accessible to all, with a strong focus on innovation in our field. What to Expect: 👉 Inspiring Talks: High-quality presentations from leading experts. 👉 Practical Insights: Feedback and case studies from experienced practitioners. 👉 Interactive Sessions: Hands-on workshops to energize and empower participants. We are thrilled to collaborate with some of the brightest minds in the ecosystem, including Christophe Blefari and Hymaïa, to curate a selective and impactful program. Conference Details: - 📅 On November 25 in Paris - Attendees are limited to 300 passionate Data & AI enthusiasts - Early Bird tickets are now available for a limited time https://lnkd.in/ext2C3fB Sponsorship Opportunities: We are also welcoming sponsors who wish to contribute to the advancement of the data and AI ecosystem. Stay updated by following our dedicated LinkedIn page for announcements: Forward Data Conference. We hope to see you there!

82

3 commentaires

#naturallanguage #atlas #charts ! Want to visualize your MongoDB data effortlessly? 📊 With Natural Language Mode in Atlas Charts, you can create sophisticated data visualizations by simply asking questions in plain English. Learn more on this latest enhancement announced at #MongoDBlocal Sydney. 👇 https://lnkd.in/gSmTbcVT

1

What are the key roles of Data Governance? 🤔 Delighted to have been invited back on DataGen at Robin Conquet's microphone to discuss. On the menu : 🧐 The key roles of Data Governance : Data Owner, Data Steward. 🚀 Current trends : Data Product Manager, Data Governance Engineer 🔥 My view on the impact of GenAI in Data Governance 📚 An article that caught my eye in recent weeks : Is Excel Immortal? Links to the episode are in comments! 👇

131

14 commentaires

Great new article on how Informatica can help customers deploy a Data Mesh. The Data Marketplace is the glue that holds it all together with consistency, governance, and security.

50

2 commentaires

💡 Meet Claude Dev: An Open-Source Autonomous AI Programmer in VS Code 💡 Claude Dev is a new autonomous AI programming assistant, leveraging the capabilities of Anthropic’s Claude 3.5 Sonnet. Developed during the Build with Claude June 2024 hackathon by Saoud Rizwan, Claude Dev is now available as an open-source extension in the VS Code Marketplace. It goes beyond simple code completion, offering comprehensive support for complex software development tasks. 📊 Key Highlights 🔹Advanced Capabilities: ◼️ Agentic Coding: Handles complex tasks step-by-step, reading and writing files, creating projects from scratch, and executing terminal commands with user permission. ◼️ Task Tracking: Monitors total tokens and API usage costs for ongoing tasks, helping manage resources effectively. 🔹Enhanced Development Workflow: ◼️ Autonomous Assistance: Provides more than just code snippets, aiding in the entire development process. ◼️ Project Management: Capable of creating and managing entire projects autonomously, significantly reducing the workload on developers. 💡 Practical Applications 🔹 Efficient Project Development: Use Claude Dev to manage the entire lifecycle of a project, from initial setup to final deployment, saving time and resources. 🔹 Complex Task Automation: Automate repetitive and complex tasks, allowing developers to focus on high-level design and problem-solving. 🔹 Resource Management: Track and optimize token usage and API costs, ensuring efficient use of resources during development. 👉 https://lnkd.in/dhnfCwSg #ClaudeDev #LLM #AIAssistedProgramming #VSCode #SoftwareDevelopment #AnthropicAI #CodingEfficiency #AIProgramming #DeveloperTools

2

Want to know more about the new 23ai JSON relational duality views ? We have updated our playlist with new 23ai videos. In the "JSON relational duality" video, I run a short demo in which I create a JSON relational duality view on top of a relational data model. Then I use this view to query and modify data using REST endpoints and GET, PUT and POST methods. I demonstrate how the database transparently maintains the relational data when I update or insert a JSON document into the view. And I also discuss about the eTag and optimistic locking mechanism. Our playlist here: https://lnkd.in/dfaWA3xf My 23ai JSON relational duality view video here: https://lnkd.in/dgbRmcSk Stay tuned, as we update regularly this playlist with new 23ai demos!

23

1 commentaire

Why Data Mesh Architecture is the Future of Data Management ? Imagine running a restaurant where each station has its own fresh ingredients tailored to their needs. That’s the essence of Data Mesh Architecture in data management. Instead of relying on a central team to handle all data, Data Mesh decentralizes it, giving each department—whether it’s marketing, finance, or HR—control over their own data. This approach addresses the bottlenecks of traditional, centralized data systems, where data processing delays can hinder timely decision-making. Problem with Traditional Data Management: In traditional setups, data flows through a central team, causing delays as departments wait for processed insights. This creates inefficiencies and slows down decision-making—a significant drawback in today’s fast-paced business environment. Data Mesh: A Game Changer Data Mesh shifts the paradigm by distributing data ownership across different domains within an organization. Each domain manages its data, allowing for real-time insights and faster decision-making. This decentralized approach is already being used by companies like Netflix and Spotify to stay agile and responsive. Key Benefits : - Speed and Scalability : Data Mesh eliminates bottlenecks, allowing companies to scale data operations smoothly. - Real-Time Insights : Departments access and analyze their data instantly, facilitating quicker, informed decisions. - Enhanced Governance: Each domain takes responsibility for its data, ensuring better governance and compliance. Challenges to Consider: Adopting Data Mesh requires a cultural shift. Teams must take ownership of their data, requiring new skills and a change in mindset. Clear governance policies are essential to maintain data integrity and security. Why Data Mesh is the Future : As data volumes grow, the inefficiencies of traditional architectures become more evident. Data Mesh offers a scalable, flexible solution, enabling organizations to react quickly and innovate continuously. It’s not just a new way of managing data; it’s a transformative approach that empowers businesses to thrive in a data-driven world. #DataMesh #Datastrategy #AI Sushant Rajput

32

Today we released RonDB 22.10.7, 22.10 is the stable serie of RonDB. Most of the new features were intended to make it easier to integrate RonDB into Kubernetes. This version runs in Hopsworks version 4.x that is a Kubernetes cluster. The new features were: RONDB-789: Find out memory availability in a container RONDB-785: Set LocationDomainId dynamically RONDB-784: Performance improvement for Complex Features in Go REST API Server RONDB-776: Changed hopsworks.schema to use TEXT data type More details can be found in: https://lnkd.in/gaWyY5s8

22