engblogs

Leveraging vision-language models to scale data annotation for autonomous construction, as shown by Bedrock Robotics, to accelerate deployment of physical AI systems.

Apple’s 2025 Workshop on Reasoning and Planning surveys advances in reasoning, planning, model development, and embodied multimodal AI systems shaping adaptable, trustworthy agents.

Spark Declarative Pipelines (SDP) enables end-to-end declarative data engineering by moving from manual orchestration to pipeline-level planning and execution inside Apache Spark, automating incremental processing, data quality rules, and backfills.

Firefly delivers nanosecond-precision, software-driven clock synchronization across data centers by combining layered internal NIC timing with distributed consensus on random graphs, enabling scalable, fault-tolerant timing on commodity hardware.

Hexagon accelerates AI model production by deploying Amazon SageMaker HyperPod to scale training of point-cloud AI models with high-performance GPUs, enabling faster development, deployment, and end-to-end MLOps observability.

Sonrai demonstrates an end-to-end MLOps framework on SageMaker AI to accelerate precision medicine biomarker trials through secure data management, robust experiment tracking with MLflow, and governance-ready model deployment.

GA of Zerobus Ingest, a serverless, single-sink ingestion service in Lakeflow Connect that streams data directly to Delta tables in the lakehouse, reducing cost and complexity while delivering real-time, governance-enabled analytics at scale.

An actionable guide to building agentic AI with a multi-model framework using Hugging Face smolagents on AWS, showcasing vector-enabled medical knowledge retrieval across SageMaker, Bedrock, and containerized backends.

AI-assisted, API-first compliance automation enables 24x faster audits by replacing fragile manual workflows with deterministic evidence pipelines across Salesforce's mobile environments.

Cloudflare One becomes the first SASE platform to support standards-based post-quantum encryption across SWG, Zero Trust, and WAN, delivering hybrid ML-KEM-based security for TLS, MASQUE, and IPsec in a cloud-native, interoperable solution.

A reinforcement-learning driven framework to learn per-head Key-Value cache eviction policies for LLM inference, boosting long-context efficiency without modifying the model.

A concise analysis of OpenAI's Frontier Alliance Partners announcement and its implications for AI collaboration, governance, and ecosystem partnerships.