engblogs

A technical walkthrough of how 7-Eleven replaced scattered maintenance documents with a Databricks Agent Bricks powered AI assistant, integrating Unity Catalog, vector search with Embeddings Compute, and a Teams Bot to surface contextual manuals, diagrams and images in seconds while reducing downtime.

Unifying GenAI on Databricks to deliver a safe, scalable home-services platform through a hybrid CNN-LLM workflow that enhances privacy, trust, and collaboration across Thumbtack’s data science and engineering stack.

Explores scaffolds and prototyping tools for shaping the user experiences of interface agents, via AgentBuilder design probes, requirements elicitation, and in-situ on-device multi-agent prototyping.

AdaBoN introduces a prompt-adaptive Best-of-N alignment framework featuring a two-stage algorithm: an initial exploratory phase to estimate per-prompt reward distributions, then adaptive budget allocation to improve LM-RM alignment while reducing inference latency across diverse prompts.

A concise technical overview of the cross-sector partnership between OpenAI, SoftBank Group, and SB Energy to explore AI-powered energy technologies.

Replace read-only ephemeral sandboxes with Sprites—durable, instantly bootable computers that support checkpoint/restore, ample storage, and global Anycast access for seamless development-to-prod workflows.

An end-to-end guide to accelerating LLM inference via post-training weight-and-activation quantization using AWQ and GPTQ on Amazon SageMaker AI, reducing memory and latency without retraining.

Introducing the Snorkel Agentic Coding Benchmark—a real-world, end-to-end evaluation suite for AI coding agents that spans multiple languages, 100 multi-step tasks across four difficulty tiers, and rigorous long-horizon planning, error recovery, and sandboxed execution.

A principled taxonomy for evaluating speech models that maps evaluation types to model capabilities and task requirements, revealing gaps in prosody, interaction, and reasoning to guide benchmark design and practical model selection.

Case study: Hackensack Meridian Health de-risks a large Google Cloud network migration by using VPC Flow Logs and Flow Analyzer to visualize Cloud Interconnect traffic from VLAN attachments via Sankey diagrams.

Introducing PPGen, a biophysical model that maps photoplethysmography signals to interpretable optical and physiological tissue parameters, and Hybrid Amortized Inference (HAI) for fast, robust, and scalable parameter estimation in wearables, balancing deep learning performance with clinical interpretability and hardware-aware design.

A memory-augmented pretraining approach using hierarchical memories to separate long-tail knowledge from common knowledge, with a small anchor model accessing a large memory bank for memory-efficient inference and fine-tuning on resource-constrained devices.