Apache Airflow 2 officially reached end of life on April 22, 2026, meaning no more security patches, bug fixes, or provider updates. Teams still running Airflow 2 now face increasing risks, especially as dependencies evolve and compatibility drops. The shift to Airflow 3 introduces architectural changes and requires careful migration planning.
Pinterest shares how it handles large-scale URL normalization to improve content deduplication. The system reduces redundancy, improves data quality, and ensures consistent indexing across billions of URLs, highlighting the importance of preprocessing in large data systems.
A structured overview of multi-agent systems, covering architecture, coordination patterns, and real-world applications. Agentic systems are shifting AI from passive tools to autonomous systems that plan, act, and collaborate to achieve goals.
Agentic AI is expected to transform payment systems by automating decision-making, fraud detection, and transaction flows. The paper explores both efficiency gains and risks, including governance, security, and regulatory implications.
Anthropic shares a detailed postmortem of a recent system incident, highlighting failure modes in large-scale AI infrastructure. The report focuses on reliability, monitoring, and lessons learned from production outages.
OpenAI introduces an open-source privacy filter designed to remove sensitive information directly on-device. This approach reduces reliance on centralized processing and improves data protection for enterprise AI workflows.
A curated overview of the latest AI research papers, highlighting emerging trends across LLMs, multimodal systems, and agentic architectures.
Google announces updates across Gemini models, agentic AI capabilities, and TPU infrastructure. The focus is on scaling AI systems for enterprise deployment and integrating agents across cloud services.
A talk on why AI coding tools don’t replace engineering discipline. The key difference isn’t the tool but the process: developers who succeed use structured approaches like vertical slices, TDD, and shared language. The message is clear — classic software engineering principles matter even more when working with AI.
A full hands-on workshop covering the lifecycle of AI-assisted development — from turning vague ideas into structured PRDs to running autonomous coding agents. It demonstrates how to combine human-in-the-loop workflows with fully autonomous runs, and how to design codebases that maximize agent effectiveness.
A large open-weight model available through Ollama, designed for local deployment and experimentation. It supports advanced reasoning and can be integrated into agent workflows and local AI systems.
A session focused on the importance of data lineage in modern AI platforms, covering governance, observability, and trust in data pipelines.
A major event focused on cloud-native architectures and enterprise AI. Topics include platform engineering, AI infrastructure, and building scalable systems across modern cloud environments.
A regional conference bringing together data engineers, analysts, and AI practitioners. Covers modern data platforms, analytics, and practical AI implementations across industries.