Data Science Dojo

Want to be a pioneer in AI? Our immersive bootcamp is your gateway to mastering LLMs and building the future of enterprise applications. ⏰ Last Bootcamp of the year is here – Secure your spot now ➡️ datasciencedojo.com/bootcamps/llm-… Join us for 5 days of: ✔ Hands-On Learning: Build your own LLM projects with expert guidance. ✔ Industry Insights: Learn from top AI professionals and gain practical knowledge. ✔ Career-Boosting Skills: Acquire the skills that are in high demand in the AI world. Our comprehensive curriculum covers: 🔸LLM and generative AI landscape 🔸Barriers to enterprise adoption of generative AI 🔸Risks and challenges in building LLM applications 🔸Embeddings, attention mechanism and transformer architecture 🔸A practical introduction to vector databases 🔸Building LLM applications with LangChain 🔸Fine-tuning and deploying large language models 🔸LLM observability and monitoring 🔸Evaluation of LLM applications 🔸Guardrails and responsible AI 🔸Challenges in building RAG applications 🔸Domain and task-specific LLMs 🔸Productionizing LLM applications 🔸Capstone project: Building and deploying an LLM application using GitHub and Streamlit #llm #llmdojo #machinelearning #ai #generativeai #finetuning #langchain #llmapplications #RAG #vectordatabases #embeddings

🚀 Excited to have Chris McCormick, AI Researcher & Educator, lead a hands-on workshop at the Future of Data and AI: Agentic AI Conference — 12:00–2:00 PM Pacific! In "Hands-On GRPO Training: From GRPO Theory to Practice", Chris moves beyond theory and equations to actually running a stable GRPO training loop — fine-tuning a Qwen model on arithmetic tasks with a focus on the design decisions that determine whether reinforcement learning trains successfully. In this workshop, you'll learn to: - Walk through a full GRPO training pipeline covering generation, reward calculation, and optimization - Understand the key hyperparameters behind stable RL training — prompt batch size, group size, learning rate, and dataset difficulty - Interpret RL metrics like the standard deviation of rewards to evaluate training health - Diagnose instability and reward hacking using Weights & Biases - Scale GRPO systems using tools like vLLM and multi-GPU training 🎟️ Save your spot: hubs.la/Q047yrCp0 #agenticai #aiconference #datasciencedojo #grpo #rlhf #llms #machinelearning #aiworkshop

Attention is all you need. But do you actually know why? We're bringing Luis Serrano (Founder & CEO, Serrano Academy) to our Large Language Models Bootcamp for a power-packed session on Transformer Architecture & Attention Mechanisms. Here's what you'll walk away understanding: 🔹 Why most developers can't explain what their LLM is actually doing — and how to fix that 🔹 The architectural decisions behind transformers that determine when they succeed and when they fail 🔹 How attention mechanisms decide what matters — and what gets ignored 🔹 Why embeddings are the foundation of retrieval, reasoning, and search — and how to use them 🔹 Where fine-tuning goes wrong, and how to avoid the most common pitfalls 🔹 How to ground LLM outputs in real data using semantic search And it's not just theory. You'll get hands-on exercises with Sentence Transformers, semantic search, and attention scoring so you leave with skills, not just slides. 📅 March 23rd, 2026 Secure your spot at the LLM Bootcamp: hubs.la/Q047CPpb0 This is just one session inside our full LLM Bootcamp — a complete learning experience designed to take you from the fundamentals all the way to building with LLMs. If you're serious about understanding and working with AI, this is where you start. #llmbootcamp #largelanguagemodels #transformer #aibootcamp #generativeai #attentionisallyouneed

Join João Moura, David Park, Bob van Luijt, and Raja Iqbal at the Future of Data and AI: Agentic AI Conference for the panel "From Hype to Durable Value: The Economics and Enterprise Reality of Agentic AI" — April 6, 10:15–11:00 AM Pacific! Agentic AI sits at the intersection of hype, capital intensity, and organizational redesign. This panel tackles the real questions executives are facing — why AI deployments stall, what separates sustainable value from speculative momentum, and how to design human-agent collaboration models that enhance performance rather than degrade it. A must-attend strategic discussion for leaders deciding where autonomy truly creates long-term enterprise advantage. 🎟️ Save your spot: hubs.la/Q047rQ8S0 #agenticai #futureofdataandai #dataandai #aiconference #datasciencedojo #enterpriseai #aistrategy #agenticsystems #aieconomics #futureofwork

🚀 Excited to have Ankit Khare, Developer Relations Specialist at LandingAI, lead a hands-on workshop at the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! In "The Last Mile of OCR/LLM-Based Document AI: Hands-On with Agentic Document Extraction", Ankit tackles the real challenge that benchmarks don't capture — large tables, old scans, mixed-language docs, handwriting, and complex layouts and shows how LandingAI's Agentic Document Extraction goes beyond OCR and parsing to handle enterprise document workloads that actually exist in the wild. In this workshop, you'll learn to: - Understand the pillars of Agentic Document Extraction and where traditional OCR falls short - Build document processing pipelines using the ADE API and SDK - Use Skills to have coding agents build document workflows for you - Analyze LLM performance on large tables, scanned docs, and complex layouts - Enable LLMs with structured output from ADE to close the last-mile gap 🎟️ Register for the workshop: hubs.la/Q047rBtB0 #agenticai #aiconference #datasciencedojo #landingai #documentai #agenticdocumentextraction #ocr #enterpriseai

🚀 Excited to have Andrea Kropp, Applied AI Engineer at LandingAI, lead a hands-on tutorial at the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! In "Agentic Document Extraction at Scale: Building a Self-Improving Pipeline with Multi-Agent Orchestration", Andrea walks through why brittle OCR pipelines and one-shot VLM prompts break at scale — and presents a production architecture that measures its own accuracy, identifies what's failing, and fixes itself automatically. In this tutorial, you'll learn to: - Build a document processing pipeline from raw PDF through the Parse API to grounded markdown - Use the Extract API to generate structured field output from complex, high-volume documents - Design an orchestration layer that routes documents to the right schema automatically - Score extractions against a golden eval set and drive targeted refinement based on evidence - Walk away with a fully replicable blueprint — all code and configuration included 🎟️ Save your spot: hubs.la/Q047rCWR0 #agenticai #aiconference #datasciencedojo #landingai #documentai #agenticdocumentextraction #multiagentsystems #ocr

🚀 From Zero to AI Agent: A Practical ReAct Build Session ReAct agents combine reasoning (thinking through problems) with actions (using tools and APIs), making them far more capable than traditional prompt-based systems. Curious how real AI agents are actually created - not just talked about? Step into a live, hands-on workshop with Kwasi Ankomah, Lead AI Architect at SambaNova Systems, and learn how to design and implement a ReAct-style AI agent using LangGraph and MiniMax-M2.5. 📅 25 March 2026 | 11:00 AM Register Now: hubs.la/Q047zr590 You’ll learn how to: • Build the ReAct loop (Reason → Act → Observe) • Integrate tools and commands into agents • Manage state and context across steps • Design reliable AI agents By the end, you'll have a working AI agent with a TODO planner and virtual file system. #AgenticAI #LangGraph #AIAgents #AIEngineering #SambaNova

Join Reena Agarwal, Tushar Jain, Philip Rathle, and Muazma Zahid at the Future of Data and AI: Agentic AI Conference for the panel "Governing Autonomy: Policy, Control, and Accountability in Agentic AI Systems" — April 6, 9:15–10:00 AM Pacific! As agentic AI systems gain the ability to reason, act, and collaborate across tools and environments, governance can no longer be an afterthought. This panel brings together experts to explore how enterprises can scale autonomy responsibly, through architectural control mechanisms, human-in-the-loop oversight, regulatory accountability, and robust safety design, before enforcement or operational failure forces reactive redesigns. A must-attend conversation for every enterprise leader navigating the realities of production agentic AI. 🎟️ Register for free: hubs.la/Q047rJMn0 #agenticai #futureofdataandai #dataandai #aiconference #datasciencedojo #aigovernance #enterpriseai #agenticsystems #responsibleai

AI job listings in 2026 expect a lot more than prompt engineering. The 10 skills in this carousel are what companies are actually hiring for right now. Swipe through and see where you stand. Our LLM Bootcamp covers all of these — hands-on, not just theory — so you're building the kind of experience that shows up on a resume and holds up in an interview. Which of these are you weakest on? Drop it below, and if you're ready to close the gap, register now for the LLM Bootcamp happening 23rd March 2026. Link to register in the comments 👇 #LLM #AIEngineering #AgenticAI #DataScienceDojo #LLMBootcamp

Joining us at the Future of Data and AI: Agentic AI Conference, April 6–10, 2026 — meet Muazma Zahid, Group Product Manager at Google BigQuery! ⭐ Muazma will be on the panel "Governing Autonomy: Policy, Control, and Accountability in Agentic AI Systems", exploring how enterprises can scale autonomy responsibly through architectural control mechanisms, human-in-the-loop oversight, regulatory accountability, and robust safety design, before enforcement or operational failure forces reactive redesigns. Muazma shapes the future of data and AI workloads at Google Cloud, with a background leading product strategy for Azure SQL Databases at Microsoft — driving innovations in vector search, AI-ready database capabilities, and intelligent data applications. A researcher in Biomedical Engineering with international publications, and a longtime advocate for diversity in tech through roles at AnitaB org, WomenWhoCode, and Women@Microsoft, she brings both deep technical expertise and a commitment to making AI accessible and impactful. 🎟️ Reserve your spot: hubs.la/Q047qrWC0 #agenticai #aiconference #datasciencedojo #speakerSpotlight #google #bigquery #aigovernance #enterpriseai #womeninai

🚀 Get ready for hands-on learning at the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! We're excited to announce our full workshop lineup, featuring sessions led by Docker, LandingAI, Serrano Academy, Contextual AI, and more! 1️⃣ From Isolation to Trust: Building Secure Sandboxes for Autonomous AI Agents by Oleg Šelajev, AI & Developer Relations at Docker: Build a workflow for running autonomous agents you can actually trust — without the security nightmares. 2️⃣ Reinforcement Learning with Human Feedback and GRPO: How Do You Get Your LLM to Do Math by Luis Serrano, Founder of Serrano Academy: Understand why LLMs needed a fundamentally different training approach for math and coding — and implement it with GRPO. 3️⃣ From Retrieval to Orchestration: Building Intent-Driven Agentic Context Engineering Systems by Scott Askinosie, Contextual AI: Build a full multi-agent context engineering system from the ground up — vector databases, query decomposition, intent orchestration, and evaluation included. 4️⃣ The Last Mile of OCR/LLM-Based Document AI: Hands-On with Agentic Document Extraction by Ankit Khare, Developer Relations at LandingAI: Go beyond OCR and parsing to tackle the real-world document AI challenges that even the best benchmarked models still struggle with. 5️⃣ Hands-On GRPO Training: From GRPO Theory to Practice by Chris McCormick, Independent AI Researcher & Author: Move from theory to a running GRPO training loop — fine-tuning a Qwen model on arithmetic tasks with a focus on the decisions that make RL training actually work. 🎟️ Register for all workshops here: hubs.la/Q047q6xT0 #agenticai #aiconference #aiworkshop #handsonai #docker #rlhf #grpo #rag #documentai #landingai #multiagentsystems

⚡ A real AI agent isn’t just a model - it’s a system that can reason, act, and improve with every step. This infographic breaks down the core pieces behind a ReAct agent architecture: • Agent Loop — Reason → Act → Observe • Tools — File system, TODO planner, command execution • State Management — Maintain context across steps • Structured Outputs — Reliable tool inputs and outputs • Multi-Step Reasoning — Repeat until the task is complete Together, these components turn a simple LLM into an AI agent that can plan and execute real tasks. In this hands-on session, Kwasi Ankomah, Lead AI Architect at SambaNova Systems, will walk through how to build a ReAct AI agent using LangGraph and MiniMax-M2.5 - from the core agent loop to tool integration and state management. 🚀 📅 25 March 2026 | 11:00 AM Register Now: hubs.la/Q047s04g0 In this session you'll learn how to: • Implement the ReAct loop (Reason → Act → Observe) • Connect tools and commands to your agent • Manage context and state across steps • Design AI agents that are reliable in real workflows You'll leave with a working AI agent you can build upon. #AgenticAI #LangGraph #AIAgents #AIEngineering #SambaNova

We're excited to welcome Chris McCormick, Independent AI Researcher & Author, to the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! 🌟 ⭐ Chris will be leading the hands-on workshop "Hands-On GRPO Training: From GRPO Theory to Practice", a practical companion to Luis Serrano's RLHF session that moves from theory and equations to actually running a stable GRPO training loop, fine-tuning a Qwen model on arithmetic tasks with a focus on the design decisions that make RL training work. Chris is an independent AI researcher and educator best known for his widely-read blog at hubs.la/Q047q0RF0, where he breaks down the inner workings of transformer models with clarity and depth. His recent research series, Patterns and Messages, re-frames attention using merged matrices to explore low-rank structure and efficiency trade-offs — reflecting his deep focus on transformer architecture, interpretability, and making complex AI research genuinely accessible. 🎟️ Register for the workshop: hubs.la/Q047q4jm0 #agenticai #aiconference #datasciencedojo #speakerSpotlight #grpo #rlhf #llms #transformers #airesearch

We're thrilled to welcome Philip Rathle, CTO at Neo4j, to the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! ⭐ Philip will be on the panel "Governing Autonomy: Policy, Control, and Accountability in Agentic AI Systems" — exploring how enterprises can scale autonomy responsibly through architectural control mechanisms, human-in-the-loop oversight, regulatory accountability, and robust safety design, before enforcement or operational failure forces reactive redesigns. Philip is a product and technology executive with 25+ years in enterprise data, currently serving as CTO at Neo4j. As an early architect of the graph database category, he oversaw Neo4j's product transformation from a single on-prem database to a cloud-based portfolio exceeding $100M ARR — and brings that same depth of systems thinking to the challenges of production-grade agentic AI. 🎟️ Reserve your spot now: hubs.la/Q047n6Zs0 #agenticai #aiconference #datasciencedojo #speakerSpotlight #neo4j #graphdatabases #enterpriseai #aigovernance #agenticsystems

🚀 Excited to have Mahdi Ghodsi and Eda Zhou from AMD lead a hands-on tutorial at the Future of Data and AI: Agentic AI Conference — April 7, 10:10–11:00 AM Pacific! In "Self-Hosted AI Agents: Running Open-Weight Models on AMD GPUs", Mahdi and Eda walk through how to build your own personal AI agent with full control over your stack — reducing API costs and running open-weight models on AMD GPUs using modern agentic frameworks like OpenClaw. In this tutorial, you'll learn to: - Understand the core concepts behind modern agentic frameworks and open-weight model hosting - Set up and run open-weight models on AMD GPUs for private, cost-efficient inference - Assemble a tool-using agent that's customizable and ready for real workflows - Reduce dependency on third-party APIs while maintaining full control over your agent stack 🎟️ Save your spot: hubs.la/Q047f7FK0 #AgenticAI #AIConference #AMD #SelfHostedAI #OpenWeightModels #AIAgents #LLMs #AITutorial

We're excited to welcome Bob van Luijt, Co-Founder & CEO of Weaviate, to the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! Bob will be joining the panel "From Hype to Durable Value: The Economics and Enterprise Reality of Agentic AI", examining why agentic AI projects stall, what separates sustainable value from speculative momentum, and how organizations can design for durable enterprise transformation rather than speculative momentum. Bob van Luijt is a technology entrepreneur, technologist, and new media artist from the Netherlands. He is the co-founder of Weaviate and the chairman of the Creative Software Foundation. In March 2016, Van Luijt started the open source vector search engine Weaviate. He has published and lectured about (open-source) software business models and the positioning of broadly applicable infrastructure software (e.g., databases and search engines). 🎟️ Save your seat: hubs.la/Q047f4fv0 #agenticai #futureofdataandai #aiconference #datasciencedojo #SpeakerSpotlight #weaviate #vectordatabases #enterpriseai #agenticsystems

⚡Simple chatbots can answer questions but they can't think through a problem, take action, or remember what just happened. That’s why we need to move beyond chatbots and start building real agents. This infographic shows how a ReAct agent solves this problem: A user gives an input → LangGraph manages the workflow → the AI model decides what to do → the agent uses tools → memory keeps track of context → and the agent returns the final result. Join Kwasi Ankomah, Lead AI Architect at SambaNova Systems, for a hands-on session where you'll build your own ReAct AI agent using LangGraph and MiniMax-M2.5. 📅 25 March 2026 | 11:00 AM Register Now: hubs.la/Q047f4170 In this session you'll learn how to: • Implement the ReAct loop (Reason → Act → Observe) • Connect tools and commands to your agent • Manage context and state across steps • Design AI agents that are reliable in real workflows You'll leave with a working AI agent you can build upon. #AgenticAI #LangGraph #AIAgents #AIEngineering #SambaNova

🚀 Get ready for hands-on learning at the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! We're excited to announce our tutorial lineup, featuring sessions led by Microsoft, Docker, Google DeepMind, AMD, and LandingAI! 🎯 Agentic Document Extraction at Scale: Building a Self-Improving Pipeline with Multi-Agent Orchestration by Andrea Kropp, Applied AI Engineer at LandingAI: Build a production pipeline that measures its own accuracy, identifies what's failing, and fixes itself — from raw PDF to structured output. 🎯 GitHub Copilot Everywhere: CLI, VS Code, and the Cloud by Kayla Cinnamon, Senior Developer Advocate at Microsoft: Build a real feature end-to-end using GitHub Copilot across the full development loop — and learn honestly where each surface shines and where it doesn't. 🎯 Securing AI Coding Agents with Docker Sandboxes and the MCP Toolkit by Michael Irwin, Principal Software Engineer at Docker: Explore real attack scenarios and the guardrails Docker is building to give agents full power — safely. 🎯 Self-Hosted AI Agents: Running Open-Weight Models on AMD GPUs by Mahdi Ghodsi & Eda Zhou, AMD: Build your own personal AI agent with full control, reduced API costs, and open-weight models running on AMD hardware. 🎯 Antigravity and AI Studio with the Gemini APIs: Building and Deploying AI Applications with Google's Developer Toolkit by Paige Bailey, AI Developer Relations Lead at Google DeepMind: Go from idea to production faster using Google AI Studio and the Gemini APIs — multimodal, long context, and ready to build. 🎟️ Register now: hubs.la/Q0477-2s0 #agenticaiconference #datasciencedojo #microsoft #docker #googledeepmind #landingai #amd

📢 Kimi AI just released a paper showing you can match the performance of a model trained with 1.25x more compute by changing one thing: how residual connections work. The core problem is something that has been sitting inside every transformer since 2015. When layer outputs accumulate through a network, every layer gets the same fixed weight of 1. By layer 50, earlier layers are contributing so little to the final result that research has shown you can remove a significant fraction of them entirely with barely any performance drop. The model had already learned to ignore them. Attention residuals replace that fixed accumulation with a learned weighted sum over all previous layer outputs. Each layer computes a small search query, scores every earlier layer's output for relevance, and builds its input from the most useful ones. The weights adapt per input rather than staying fixed, which is what makes the difference. Tested on a 48B parameter model trained on 1.4T tokens, the gains hold across every benchmark. GPQA-Diamond up 7.5 points. Math up 3.6. HumanEval up 3.1. The largest improvements are on multi-step reasoning tasks, which makes sense — those are exactly the tasks where later layers need to selectively build on what earlier layers figured out. Full breakdown in the blog. Link in the replies! #AttentionResiduals #KimiAI #LLM #DeepLearning #AIResearch #GenerativeAI #DataScience

We're excited to welcome Andrea Kropp, Applied AI Engineer at LandingAI, to the Future of Data and AI: Agentic AI Conference, April 6–10, 2026! Andrea will be leading a hands-on tutorial on "Agentic Document Extraction at Scale: Building a Self-Improving Pipeline with Multi-Agent Orchestration" — walking through a production architecture that measures its own accuracy, identifies what's failing, and fixes itself, taking documents from raw PDF all the way through to structured field output via LandingAI's Agentic Document Extraction API. Andrea brings 15 years of applied data science and ML experience, having helped enterprises productionize agentic document extraction and computer vision at LandingAI. A former Senior Director of Data Science at a Fortune 500 company, she holds an M.S. in Physical Chemistry from the University of Michigan. 🎟️ Grab your spot: hubs.la/Q046_vy30 #AgenticAI #AIConference #LandingAI #DocumentAI #AgenticDocumentExtraction #MultiAgentSystems #EnterpriseAI