Application team tech lead on Cloud SIEM product. Responsible for the overall API, telemetry parsing/enrichment, and aggregating telemetry into users, domains, etc. Architected and built the threat intelligence collection and enrichment capabilities over the course of a year. Sustained the effort to integrate Cloud SIEM into the core product, as well as reduce cloud costs and maintain SLAs.Technologies: Kafka, Kubernetes, Flask, Terraform, AWSLanguages: Scala, Python

Greenfield creation of the Nucleus (anaconda.cloud) site, building infrastructure, microservices, and APIs to support its growth from 0 to 2+ million users.Technologies: Kubernetes, Istio, AWS, Postgres, FastAPI, TerraformLanguages: Python, Golang

Building and maintaining the Anaconda Enterprise 5 platform, which provides data scientists and data engineers with workflows for collaboration and deploying code & notebooks.Technologies: Kubernetes, JupyterLab, TornadoLanguages: Python

Lead data scientist for the Endpoint Protection Product team, combining machine learning knowledge with high performance cybersecurity product experience to build resilient malware detection capabilities. Built highly scalable machine-learning infrastructure (100TB+ sample set sizes) for rapid experiment iterations and broad threat coverage. Provided evidence and model building transparency for leadership, from raw data gathering to 3rd party efficacy tests. I also built the primary detection and prediction capabilities in the DeepArmor endpoint client.Technologies: Apache Spark, XGBoost, SKLearn, TensorFlow, .NETLanguages: Python, C#

Designing and building infrastructure and features for SparkSecure, an automated solution which detects cybersecurity anomalies with machine learning and anomaly validation using natural language processing techniques across various knowledge bases. Furthered development of NLP techniques applied to automated threat research and enabled greater scalability on data ingestion and anomaly detection. We discontinued this product as a whole in order to focus on its ML-based endpoint protection capabilities.Technologies: Apache Spark, Hadoop, KafkaLanguages: Scala

Lead software engineer on the Platform team, designing and implementing features for high throughput network/security event processing in a multi-process architecture, with heavy focus on performance and scalability. Also involved with performance-tuning Vertica DB to store and retrieve large volumes (terabytes) of network event data in small and medium sized multi-tenant clusters.Technologies: ZeroMQ, GLib, Linux, Vertica, MongoDBLanguages: C, Python

Part of a small and agile software company, developing products to function within cross-domain security environments for various government organizations in the intelligence community.Tasks and responsibilities included:- Adding core features to a high-performance cross-domain messaging service.- Part of the TSABI PL4 certification effort for the cross-domain messaging service.- Designing and implementing an installation and administration front-end for the product line.- Researching and recommending new technologies.- Improving and adding features to the product line.- Developing prototypes for customer demonstrations.Frameworks & platforms: Spring, Camel, GWT, Errai, JAX-WS, Linux, Solaris TXLanguages: Java, C, Perl

Designed and implemented I/O services for the SU (Situation Understanding) component of the Army's BCTM (Brigade Combat Team Modernization's) B-Kit installation. These two services provided a large portion of battlespace data I/O for SU, and were designed to be high throughput in addition to providing functions such as data validation, path down-sampling, and transaction consolidation. The team functioned on a highly organized development schedule, requiring nearly code-accurate UML diagramming, preliminary and detailed design documents, formal code reviews, and other rigorous software processes.Platforms, frameworks, & languages: Linux, Boost, C++, Perl, UML

Developed a plugin for Open|SpeedShop to trace and aggregate system call statistics across multiple nodes on LANL's various HPC environments. Also profiled MRNet performance on the RoadRunner supercomputer, and supported general Open|SpeedShop development and documentation tasks.Platforms & languages: Linux, C++, Python, Qt

Developed a graphical front end using Tcl/Tk to run third party optical simulation software and display the results in SciLab. The resulting software suite provided photonic bandgap configuration characterization for phase shift mask lithography.