Featured Issues

How to boost LLM performance during pre-training: A preview AI_Distilled #97: What’s New in AI This Week Build Your AI Chatbot with Free LLM Boomcamp Join LLM Zoomcamp, a free online course starting on June 2 and build an end-to-end AI chatbot tailored to your use case. In 10 weeks, you’ll learn key skills like working with LLMs and RAG, vector search for indexing and retrieval, how to evaluate and monitor performance, and key best practices for building robust, real-world applications. REGISTER NOW FOR FREE It’s time for the final issue of May 2025. In this edition, we bring you the top five news highlights of the week, upcoming events shaping the AI and LLM landscape, and a sneak peek into techniques for optimizing LLM performance. LLM Expert Insights, Packt In today's issue: 🧠 Expert Deep Dive: This week, we explore pre-training optimization techniques—from quantization to flash attention—for building faster, smarter LLMs. 📅 Webinar Watchlist: June’s top AI/LLM webinars cover automation, cybersecurity, healthcare, legal AI, and multimodal fine-tuning. 🔌 Build AI Agents This Weekend: Join Packt’s Accelerated Agentic AI Bootcamp—hands-on, fast-paced, and 35% off. 📚 Optimize Your LLM Stack: Learn more from Generative AI with Python and PyTorch—a guide to efficient training and deployment. 🚀 DeepSeek V3 Debuts: China’s latest open-source model steps up with better reasoning and dev capabilities. 📰 Publishers vs. AI Search: Google CEO Sundar Pichai defends AI-powered results amid growing backlash from content creators. 📱 Apple Rebrands for 2026: WWDC will unveil iOS 26 and align all platforms under a unified OS naming strategy. 🎨 Sam Altman x Jony Ive: OpenAI teams up with the design legend to build magical, AI-first consumer products. 🧠 Anthropic Traces Thoughts: Claude’s internal reasoning gets visualized through groundbreaking interpretability research. 📈UPCOMING EVENTS JUNE'S MUST ATTEND AI/LLM WEBINARS In June 2025, a number of exciting AI webinars are already generating buzz. Here are the Top 5 not-to-miss events in the next month (for more information and registration details, please visit the links): 1. AI-Enhanced Motion Control: Innovations Driving Automation Forward Date: June 5, 2025 Time: 12:00 PM – 1:00 PM ET Location: Online Cost: Free Hosted by the Association for Advancing Automation, this webinar explores how AI is revolutionizing motion control systems, enhancing precision, efficiency, and adaptability across various industries. 2. AI Security Webinar – Practical Measures to Mitigate AI and Cybersecurity Risks Date: June 11, 2025 Time: 11:00 AM – 12:30 PM BST Location: Online Cost: Free Presented by The Alan Turing Institute, this interactive webinar brings together industry experts and SMEs to share practical, cost-efficient, and high-impact security measures that deliver maximum AI and cybersecurity protection for businesses. 3. Clinical Large Language Models in Healthcare – Applications, Challenges, and Opportunities Date: June 12, 2025 Time: 10:00 AM – 11:00 AM CEST Location: Online Cost: Free Organized by the Helmholtz Information & Data Science Academy in collaboration with NORA, this webinar features Anne Torill Nordsletta discussing the role of large language models in healthcare, exploring applications, challenges, and future opportunities in the clinical setting. 4. Inside the TBI Playbook: How I Use AI to Win the Hardest Cases Date: June 17, 2025 Time: 1:00 PM – 2:30 PM EST Location: Online Cost: Free Hosted by Anytime AI™, this CLE-accredited webinar features attorney Taylor Ernst sharing insights on leveraging AI in traumatic brain injury litigation. Attendees will learn about practical applications of AI tools in complex legal cases. 5. Multi-Modal LLM Fine-Tuning of Unstructured Data with Dataloop & SingleStore Date: June 18, 2025 Time: 10:00 AM – 11:00 AM PST Location: Online Cost: Free Presented by SingleStore, this webinar explores techniques for fine-tuning multi-modal large language models on unstructured data, covering integration strategies with Dataloop and SingleStore platforms. Machine Learning Summit 2025 JULY 16–18 | LIVE (VIRTUAL) 20+ ML Experts | 20+ Sessions | 3 Days of Practical Machine Learning and 35% OFF BOOK NOW AND SAVE 35% Use Code EMAIL35 at checkout when purchasing the 3-day ticket Limited to the first 50 customers EXPERT INSIGHTS PRE-TRAINING OPTIMIZATION TECHNIQUES FOR LLMs The scale of data and computation required for large language models (LLMs), along with the significant capital investment needed to train and deploy them, necessitates the exploration of optimization techniques throughout the LLM lifecycle. In this issue, we focus on potential improvements during the pre-training phase, as this is the most resource-intensive step, involving a vast amount of data and sensitivity to architectural design. Here are some techniques you can employ to improve LLM performance and efficiency: 1. Quantization: Quantization aims to reduce the number of bits needed to store these weights by binning floating-point values into lower-precision buckets. This reduces memory usage with minimal impact on performance. Small precision losses are acceptable as long as the model’s performance is within the required levels. For instance, a weight value like 3.1457898 could be quantized to 3.1458 using a scheme that retains four decimal places. Such a scheme might lead to slight changes (during the backward pass of the training step, for example, a higher margin of error) while computing loss or while updating weights. Take, for instance, 4-bit quantization, which uses small bins where the density of weights is higher and fewer larger bins for weights away from the mean. The 4-bit float representation employs an intelligent approach based on the distribution of model weights. Most weights tend to cluster near zero, with minor differences requiring higher precision, while fewer weights have larger values. To accommodate this, asymmetric binning is used: smaller bins are allocated for values near the mean to maintain precision, while fewer larger bins handle outliers further from the mean. 2. Mixed precision: This is another technique to reduce memory and computational demands without sacrificing significant accuracy. These methods combine different numerical formats, such as float16, int8, and more, to optimize efficiency and performance during training or inference. 3. Data efficiency: Large datasets are costly to process, and redundant or noisy data can negatively impact model performance. Therefore, data efficiency techniques can be applied to achieve high model accuracy and generalization with a reduced or optimized dataset. This process includes filtering data for quality, reducing redundancy, and applying sampling techniques to emphasize high-value samples. 4. Sparse attention: Instead of computing attention weights for every pair of tokens in the input sequence, sparse attention focuses only on a subset of tokens, exploiting patterns in the data or task-specific properties. To put things into perspective, think about decoder-only architectures like GPT trained with an auto-regressive language objective. Such an objective puts a constraint on the attention layer to be causal, and thus, only the lower triangular attention matrix is useful (but the computation is still done for the whole matrix). Different architectures leverage specific patterns, like local or strided attention mechanisms, to bring in efficiency in computation time. 5. Flash attention: Flash attention takes the route of hardware-based improvements and efficiencies to compute attention scores. There are two popular techniques for sparse attention: Kernel fusion and Tiling. Kernel fusion reduces the number of I/O operations by combining all steps (elementwise operations, matrix multiplication, softmax, etc.) into a single read-write operation. This technique is pretty effective during inference. Tiling, on the other hand, breaks down the overall attention calculation into smaller and manageable groups of operations that fit into fast and low-latency GPU memory. For instance, instead of computing softmax across the entire attention matrix at once, FlashAttention computes it over smaller chunks in a numerically stable and tiled fashion, thus making use of faster memory without the need to store a large matrix. 6. Mixture of Experts (MoE) architecture: MOE is an advanced architecture designed to leverage a subset of components (or experts) rather than the whole architecture itself, thereby achieving higher scalability and efficiency. The Experts in this architecture are independent modules or blocks of the network, where each can be trained to specialize in a specific task. While the Router is a module that learns to select which experts to leverage (or activate) for a given input based on different criteria. The Router itself can be a neural network. 7. Efficient architectures: There are a number of different patterns and techniques that have been developed and leveraged by different architectural improvements over the years. Some of the popular architectures are Linformer, Reformer, and Big Bird. Apart from pre-training optimizations, there are other techniques as well, such as fine-tuning and improvements in inference time. More recently, the availability and popularity of small language models and specialized hardware and frameworks has also contributed to significant improvements in the overall efficiency of resource-constrained environments. Liked the Insights? Want to dig in deeper? If you wish to learn more about these techniques or wish to dive deep into foundational aspects of the LLM ecosystem, you can check out the book, Generative AI with Python and PyTorch, Second Edition, by Joseph Babcock and Raghav Bali. BUY NOW 📈LATEST DEVELOPMENT Let’s kick things off with the top stories of the week. China is aiming for the top spot in the AI race with DeepSeek V3's latest release DeepSeek just released -V3-0324, claiming a major boost in reasoning, front-end development capabilities, and smarter tool use. The release positions DeepSeek as a serious contender to models like Code Llama and Codex. You can try out the open-source weights from this HuggingFace card. Publishers claim AI-Search is an internet takeover, Pichai defends it as an innovation In a podcast with Nilay Patel (Editor-in-Chief of The Verge), Google CEO Sundar Pichai shared candid thoughts on AI’s impact on the internet. He defended AI-generated search results amid backlash, insisting they won’t kill the open web. As Google walks a tightrope between innovation and publisher outrage, Pichai expressed confidence that AI will ultimately “enhance,” not erase, human content. He dodged revenue concerns but acknowledged the risks of unchecked AI growth. Catch the full conversation here. Apple’s branding power move with iOS26 A Bloomberg report says that Apple is set to revamp its OS branding game at WWDC-2025. The rebranding will sync all platforms with the upcoming 2026 launch year, setting the stage for a unified, modernized software identity with iOS 26, macOS 26, and watchOS 26. SamA and Ive team up for AI-first products OpenAI is collaborating with design icon Jony Ive and his firm LoveFrom to craft AI-powered products. Jony Ive, Scott Cannon, Evans Hankey, and Tang Tan led io team will collaborate closely with Open AI’s research and engineering teams, with LoveFrom leading design and creative responsibilities. Their goal: to recapture the magic, creativity, and wonder of early Apple-era technology. Hear more about their vision in this video. Anthropic inching towards interpretable AI? Anthropic just cracked open the black box of AI thinking with its latest research, Tracing Thoughts. Using a novel method called dictionary learning, researchers mapped how language models like Claude internally form and organize thoughts. They uncovered thousands of hidden features that resemble abstract concepts and reasoning steps. This breakthrough gives us a glimpse into not just what AI predicts—but how it thinks. Dive into this investigative research here. 📢 If your company is interested in reaching an audience of developers and, technical professionals, and decision makers, you may want toadvertise with us. If you have any comments or feedback, just reply back to this email. Thanks for reading and have a great day! That’s a wrap for this week’s edition of AI_Distilled 🧠⚙️ We would love to know what you thought—your feedback helps us keep leveling up. 👉 Drop your rating here Thanks for reading, The AI_Distilled Team (Curated by humans. Powered by curiosity.) *{box-sizing:border-box}body{margin:0;padding:0}a[x-apple-data-detectors]{color:inherit!important;text-decoration:inherit!important}#MessageViewBody a{color:inherit;text-decoration:none}p{line-height:inherit}.desktop_hide,.desktop_hide table{mso-hide:all;display:none;max-height:0;overflow:hidden}.image_block img+div{display:none}sub,sup{font-size:75%;line-height:0}#converted-body .list_block ol,#converted-body .list_block ul,.body [class~=x_list_block] ol,.body [class~=x_list_block] ul,u+.body .list_block ol,u+.body .list_block ul{padding-left:20px} @media (max-width: 100%;display:block}.mobile_hide{min-height:0;max-height:0;max-width: 100%;overflow:hidden;font-size:0}.desktop_hide,.desktop_hide table{display:table!important;max-height:none!important}}

Discover the AI + architecture shifts powering Salesforce’s next evolution SalesforcePulse #2: Cert Updates, Slack AI, & the Builder's Toolkit Hi , 🌍 Ohana!! AI in Salesforce isn’t just about smarter suggestions, it’s about smarter action. With the rise of Agentforce, evolving certifications, and increasingly modular architecture practices, the way we build and scale on Salesforce is shifting. This issue of SalesforcePulseexplores that shift – from how the platform is changing to how careers are evolving around it. Also, in breaking (emotional) news… Salesforce partner Tottenham Hotspur has just won the Europa League 🏆. As a die-hard Manchester United fan, I’d like to formally declare this the worst timeline – but hey, at least one of our teams is winning this season. 📬 TL;DR: Scroll-Starters 🔗 Salesforce's new certification experience is here – streamlined, smarter, and more candidate-friendly 🔗 AI Teammates in Slack bring Agentforce closer to daily work 🔗 Salesforce's AI data architecture shows how 100M+ records are unified in minutes 🔗 AI-assisted Salesforce integration shows how automation is reshaping modern sales 🎧 Data Cloud podcast unpacks strategy and career impact 🧰 Architect's Toolkit video delivers actionable frameworks from Dreamforce 🎤 (Coming soon): New voices and stories from the ecosystem Let’s dive in! 📈 Ecosystem Pulse: What's Moving the Salesforce World? Salesforce Acquires Informatica for $8B: Salesforce has finalized its acquisition of Informatica for $8 billion, aiming to bolster its AI capabilities, particularly within its Agentforce platform. While this is a power play for Data Cloud and ETL, the deeper story is about AI readiness. Informatica’s metadata-driven automation and enterprise-scale integration pipelines give Salesforce the fuel it needs to make Agentforce smarter, faster, and more context-aware. What’s in it for you: For architects, consultants, and data-minded builders, this could mean fewer manual pipelines, better data quality, and more robust AI-driven automation in the near future. In other news: Salesforce’s Certification Experience Gets a Revamp: Salesforce has redesigned its certification platform to offer a smoother registration and tracking flow. Expect a cleaner UI, centralized learning paths, and easier access to resources across your credential journey. (Certification Name Changes FAQ) AI Teammates Join the Slack Channel: Salesforce is moving closer to real-time agentic work with new AI teammates in Slack. These aren't just copilots; they’re context-aware assistants that can retrieve insights, manage workflows, and support daily operations across teams. Agentforce Gets Industry-Specific with Financial Services Launch: Salesforce is extending its AI agent framework to industry use cases, starting with financial services. This verticalized rollout signals how AI teammates will need to handle regulation, data integrity, and client sensitivity. 🧠 Expert Corner: What the Community is Talking About A quick roundup of practical insights, architecture thinking, and platform design strategies from across the Salesforce ecosystem. How Salesforce Built an AI Architecture at Scale: A behind-the-scenes look at how Salesforce unified 1,000+ data sources and 100M+ records in under five minutes. A must-read for anyone thinking about scalability, automation, and system design. AI-Assisted Salesforce Integration for Sales Teams: This guide explores how AI can supercharge Salesforce integrations, from smarter forecasting to automated sales actions — a shift toward hands-free pipeline management. Design Patterns for Integration in Salesforce: A deep dive into key integration patterns: remote process invocation, data mashups, virtualized views, and more. Ideal for architects and consultants working across systems. 🎧 Inside Salesforce Data Cloud - Podcast: In this episode of the Salesforce Posse Podcast, host Francis Pindar sits down with Ian Gotts, founder and CEO of Elements.cloud. They unpack how Salesforce Data Cloud impacts planning, implementation, and long-term scalability — and what it means for your career path and design decisions. 💬 Community Invitation Know a great resource, podcast, or video that helped you architect smarter on Salesforce? Help the Ohana community! Just hit reply, and it might feature in our next Expert Corner. 🧰 Tool of the Month 🎥 Architect’s Toolkit (Dreamforce On-Demand): In this Dreamforce episode, Salesforce experts share tools, templates, and mental models they use to design at scale. Learn how seasoned architects tackle decision frameworks, solution documentation, and communication across teams. It’s a practical watch for anyone leveling up from solutioning to system thinking. 💬 Have a go-to tool or framework that helps you think, build, or communicate better? Reply to this email or DM me on LinkedIn - we’d love to feature it in a future issue. 📚 Exclusive Book Excerpt Salesforce Platform Enterprise Architecture, 4th edition 📚 Limited-time offer: Get 20% off on both eBook + paperback versions of the full title. Buy it here. In this section, we explore an excerpt from Chapter 16 of this book by Andrew Fawcett As AI becomes core to platform-level decision-making, understanding what Salesforce Einstein offers – and how it's meant to be used by different roles – is essential. This excerpt outlines the practical Einstein tools available to admins and analysts, from prediction builders to insight discovery. Salesforce provides a number of services and features, both general and specific, related to AI so as to suit different types of users, use cases, and product features. Regardless of whether you are a developer, an admin, or an end user of one of the specific clouds (such as Sales Cloud), there is likely something for you. As you would expect, the options open to developers or admins help fill in any specific gaps where Salesforce has not delivered a specific prediction need. In this section, and those that follow, we will explore these services in detail. For administrators and data analysts, these are the options: Einstein Prediction Builder:This works with the data stored in custom or standard objects to look for ways to predict a given outcome, such as the likelihood of a given customer paying on time orusing historical race data to predict a race winner! The prediction is calculated in the background and the result is stored as a Custom Field that you can define further. You can configure predictions bycompleting a short wizard-based UI under theSetup Einstein Discovery: Scans datasets that you define to discover patterns to drive information about business events (insights) or needs so that you can be proactive in resolving issues or making improvements. One insight might be a particular region that is not performing well in terms of sales, maybe due to certain products not being well promoted, consulting teams failing due to being overworked, or staffing issues due to market demand increases. Datasets can be from Salesforce (your Standard or Custom Objects) or from externally available data. For example, in Formula 1, you might use it to compare data from your competitors against your own. It also allows you to perform what-if analysis to determine whether changing certain aspects of your business would have the desired effect. Finally, as an advanced feature, it is possible to have your developers create Lightning Components that display insights directly while viewing specific records. Einstein Next Best Action: This integrates with your user experience to provide contextual actions, depending on the field values of the record they have open. For example, you could automatically display the Give Feedback action defined earlier in this book as a recommended action, once the race completes. You can also use fields populated by Einstein Prediction Builder to display actions based on predictions; for example, creating a task for the pit stop team if it appears that your team is predicted to lose an upcoming race given the current time it takes them to change tires. Finally, it is also possible to use insights returned from Einstein Discovery to determine actions displayed to the end user. Fordevelopers,these are the options: Einstein Vision REST API: This helps you integrate image and character recognition into your applications. Image recognition allows you to not only recognize (or classify) a single item in each picture, but also detect multiple items, for example, cars and people. Salesforce provides some standard models that they have built with images they have obtained, though you also have the option to provide your own. Also included is an OCR (Optical Character Recognition) service to extra character data from images or PDF files. Einstein Language REST API:Thishelps you to integrate different ways to recognize free-format text in your application. This allows you to determine whether a phrase is reflecting a good or a bad sentiment from the user, or even pick out specific aspects of a sentence, such as a place or product being referenced. Using this API, you can discover the intent of such a message from the user, for example, as a desire to book travel to a specific place or reset a password. Finally, there is an array of ever-expanding features built into Salesforce products, such asEinstein Lead and Opportunity Scoring,Einstein Opportunity and Account Scoring and Insights,andEinstein Forecasting.You can learn more aboutSales CloudandEinsteinhere:https://www.salesforce.com/products/sales-cloud/features/sales-cloud-einstein/. Buy the book at 20% discount 🔁 Found this useful? Forward it to a colleague: Refer to a friend 📅 Events in Spotlight 🔹 Agentforce World Tour & TDX London 2025 – June 11-12, 2025 | ExCeL London Join the Salesforce community for a two-day event packed with keynotes, hands-on workshops, and networking opportunities. One of Packt’s most popular authors, Akshata Sawant, will be speaking at two sessions during the event. Whether you're attending in person or following the updates online, it’s one to keep an eye on. 🔹 Salesforce Connections (CNX) – June 11 and 12, 2025 | Chicago & Salesforce+ (Live stream available globally) Join live or virtually for sessions, Agentforce demos, hands-on workshops, and interactive agent-building challenges during the Agentic Marketing Showdown. 🔹 Trailblazing Women Summit Japan – June 11, 2025| Streaming Live from Tokyo A global virtual event celebrating women in tech and community empowerment. Hear diverse perspectives and leadership stories. 🔹 Salesforce Foundations: Kickstart Your Agentforce & Service Experience – June 26, 2025 A free, virtual deep dive into Agentforce for those using Service Cloud. Live instruction on setting up and deploying your first service agent. 💬 Want us to feature a virtual Salesforce event from your region or role? Reply to this email with your suggestions – we’ll spotlight the best ones in each issue. ✍️ Share Your Story! Have you navigated a tricky architectural challenge? Designed something that balanced complexity, speed, and scale? Whether it’s a client success, a personal breakthrough, or a lesson from the trenches – we’d love to hear it. 📝 Reply to this email or DM me on LinkedIn AND we will give away a free eBook for entry submissions. 🚀 Stay on the Trail Thanks for reading the second edition of SalesforcePulse. We’re already seeing a strong response from the community. Architects, consultants, and ecosystem experts are reaching out with ideas, insights, and collaboration offers. In the next few issues, we’ll start sharing these contributions and continue bringing you practical content, community voices, and clarity through the platform noise. Until next time. Keep building smarter. 🗣️ We'd love to hear what you think of this issue. Hit Reply to this email or DM me and share your thoughts – what helped, what’s missing, or what you'd like to see next. Cheers! Rounak Kulkarni, Editor-in-chief 🔁 Found this useful? Forward it to a colleague: Refer to a friend *{box-sizing:border-box}body{margin:0;padding:0}a[x-apple-data-detectors]{color:inherit!important;text-decoration:inherit!important}#MessageViewBody a{color:inherit;text-decoration:none}p{line-height:inherit}.desktop_hide,.desktop_hide table{mso-hide:all;display:none;max-height:0;overflow:hidden}.image_block img+div{display:none}sub,sup{font-size:75%;line-height:0} @media (max-width: 100%;display:block}.mobile_hide{min-height:0;max-height:0;max-width: 100%;overflow:hidden;font-size:0}.desktop_hide,.desktop_hide table{display:table!important;max-height:none!important}}