Palo Alto Networks

From personally building PAN's first production DL model to its current reputation as an AI-first company, I played a pivotal role in successfully transforming many of PAN's cybersecurity operations with AI -- helping to catalyze PAN ahead of countless adversaries and its industry competitors.

Selected contributions:

- Published a novel adversarial learning framework called "Innocent Until Proven Guilty" to overcome a pervasive threat to ML-based malware classification.
- Published a novel algorithm for reliable and robust SSL called "Contrastive Credibility Propagation". The solution tackled an unsolved hurdle to successful ML-based DLP.
- Designed and patented a discriminative model capable of explaining its verdicts in natural language without losing discriminative power.
- Designed a JS malware classification pipeline featuring a patented architecture that extends stacked generalization by ensembling deep CNN feature extractors of multiple data representations with XGBoost. I also designed an ultra-low memory version that supports packet-by-packet streaming of inputs for use inline on a firewall.
- Designed a URL classification pipeline featuring a patented DL architecture. The IUPG-trained model extends state-of-the-art with feature source orthogonality constraints and an expanded token encoding with HMM-based randomness scores.
- Designed a DL solution for DLP. Several iterations included dataset engineering, ethical use of private data, and novel explainable AI components.
- Designed a malicious prompt classification pipeline for GenAI security featuring multi-language support and patented components designed to optimize performance with ultra-low FPR requirements.
- Led two company-wide recurring forums to discuss and share AI research.
- Countless mentees, consultations, and cross-team collaborations.

Sr. Principal Machine Learning Engineer

Principal Machine Learning Engineer

Senior Machine Learning Engineer

Machine Learning Engineer

Rochester Institute of Technology

- Research interests include computational models in theoretical neuroscience and unsupervised streaming anomaly detection.
- The grant-funded research culminated in a successful MS thesis that redefined state-of-the-art performance in an anomaly detection benchmark and introduced the "High Order Prior Belief" algorithm.

AI Scientist

Personal mathematics and computer science tutor for underprivileged students through the Collegiate Science Technology Entry Program (CSTEP).

Tutor

- Built and prepared a malicious JavaScript dataset with several million entries.
- Designed and deployed into production a new malicious JS classification pipeline that improved both performance and latency.

Data Science Intern

Florida Institute of Technology

- Discovered and described clusters of melanoma from ultra-high dimensional genomic data using a custom machine learning pipeline composed of clustering and feature selection techniques. Later published this work in Frontiers of Genetics.
- Developed a stochastic model of tumor aggressiveness to better estimate mitotic index.

Biomathematics Research Intern

Enel X

- Built a new internal database framework using Hadoop and Impala.
- Built general-purpose data fitting library code using Gauss-Newton nonlinear regression. Demonstrated the code to data scientists in the company making use of their data stores to provide new insights.

Big Data Software Engineering Intern

A curious, self-driven, visionary, and multi-talented individual passionate about cognitive science and AI. I enjoy conducting bio-inspired AI research to solve hard, real-world problems that require the cognitive capabilities observed uniquely in animal brains. I believe modeling animal intelligence’s computational principles and functional properties is the most straightforward path toward strong AI that goes beyond traditional ML and DL. In industry, I work as an MLE building AI to provide practical solutions in various domains. I am also a self-taught pianist, guitarist, banjo player, accordion player, aspiring writer, and sketch artist who has made a habit of accomplishing whatever I set my mind on.

Taplio

Brody Kutt's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Brody Kutt's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Brody Kutt's Linkedin Analytics