engblogs

Exploring the remarkable speed improvement of Mojo🔥 compared to Python - Part 1

Autonomous visual information seeking with large language models by integrating LLMs with computer vision, web search, and image search tools to achieve state-of-the-art results on visual information seeking tasks.

Exploring AI-augmented management through interdisciplinary projects at MIT Schwarzman College of Computing

Maximizing predictive task efficiency using generative AI techniques

A blogpost about neural network pruning with combinatorial optimization and its efficiency and performance compared to existing methods.

This blogpost discusses how generative AI can accelerate the time to value for predictive tasks.

This blogpost details the adoption of dbt as the standard data transformation tool at Instacart, highlighting the motivation, architecture, and solutions it offers for tooling fragmentation, lack of modularity, and performance measurement at scale.

This blogpost discusses AVA Discovery View, an internal tool at Netflix that surfaces still frames from video content to help creatives find authentic moments that represent a title's narrative themes, main characters, and visual characteristics.

An exploration of attacks on mTLS authentication, focusing on implementation vulnerabilities that can lead to user impersonation, privilege escalation, and information leakages.

Neural network pruning with combinatorial optimization: This blogpost explores an optimization-based approach for efficient pruning of pre-trained neural networks, achieving scalability and improving performance tradeoffs.

The blogpost discusses how generative AI can accelerate the time it takes to achieve value in predictive tasks.

The blogpost discusses how the new API version simplifies payment methods code and introduces the dynamic payment methods integration path.