Aditya Saligrama

A blog post detailing a Google Workspace issue encountered while orchestrating an April Fools’ Day prank among security researchers.

Paper detailing the design and implementation of Stanford's first cloud computing course (CS 40), including its curriculum, assignments, and infrastructure for resource provisioning and autograding.

A blog post discussing the balance between theoretical foundations and practical applications in computer science curricula, advocating for incorporating practical skills.

The third part of a retrospective on teaching CS 40, sharing insights and lessons learned from being the principal instructor of a new cloud computing course at Stanford.

The second part of a retrospective on teaching CS 40, detailing the custom infrastructure built for course management, student use, resource provisioning, and autograding.

The first part of a retrospective on teaching CS 40, covering the process of designing, approving, and building content for Stanford’s first-ever hands-on intro cloud computing course.

A blog post describing the process of migrating personal infrastructure to Cloudflare with no downtime, systematizing resources with Terraform infrastructure-as-code.

An exploration of Gradescope’s autograder vulnerabilities, analysis of potential impact, and the creation of Securescope for a more secure autograder configuration.

Discusses security testing of Firebase client apps, focusing on using a clever solution to grab Google OAuth tokens for signing into databases and contributing Google sign-in functionality to Baserunner.

Details how misconfigured Firebase security rules can lead to data breaches, including a story of a vulnerability found in the Fizz app at Stanford leading to deanonymization of posts and legal threats.

Describes an incident where an EternalBlue-vulnerable machine used for Stanford’s Hack Lab course was compromised, detailing the infrastructure, incident response, and how it was used for teaching.

A continuation of a personal security upgrade series, focusing on preventing evil maid attacks using disk encryption and secure boot, summarizing resources for configuration.

Paper exploring the effectiveness of ensembling with robust and non-robust features to improve natural accuracy while maintaining adversarial robustness, published in the ICLR 2020 workshop on trustworthy machine learning.

A student abstract presented at the AAAI Conference on Artificial Intelligence.

Paper analyzing how the Rust language aids developers in writing concurrent code, complementing the development of a lock-free concurrent hashmap.

A paper introducing a novel AI method for detecting polarity in online content.

Paper on systems optimizations for learning certifiably optimal rule lists, presented at the SysML Conference.