At Wind River I served as the technical lead for the AI org. I helped the org grow from 5 to 40 people. My day to day consisted of transforming feature requests from product management into technical designs, grooming stories into actionable work items, setting the quality bar in code reviews and acceptance review meetings, designing test plans, building infrastructure, and skilling up our team in our toolset and design philosophies.The AI organization was the data centric team of Wind River Studio. We built out the companies aptitude in all facets of data management from security and access controls, telemetry systems for cloud and edge, secure transmission for data, streaming and batch analytics, model training and deployment, alerting and feedback systems, and evolving database schemas.The AI organization managed several products within Wind River Studio. Our DFL (Digital Feedback Loop) product secures a trust-minimized two-way communication channel between the device and the cloud. DFL enabled applications can be invoked from the cloud without needing to open a port or webservice on the device. Telemetry from the device was used for analytics, alerting, and model training. Our OTA (Over the Air) system managed software and configuration updates for enabled IoT devices. Our PlatformHealth system managed cloud telemetry and was used for monitoring and autoscaling underlying infrastructure.I also helped build a lot of the infrastructure that goes into scaling any organization including CI/CD pipelines, environment management systems, secret management systems, network security systems, etc.

At Xevo I served as architect and backend engineer for everything from a recommendation engine doing intent classification and fulfillment in cars, to an experimentation platform using multi-armed bandit and gradual deployments, to a patent pending bandwidth reducing network proxy for connected head units using AWS IoT, to a contact tracing system to help prevent the spread of Covid-19 at factories. I authored dozens of microservices using .Net, and Java Spring Boot. I wrote batch and streaming data pipelines using SnowflakeDB, AWS Lambda, AWS RDS, and microservices. I also set many conventions for our team such as defining our test patterns, template services, and CI/CD pipelines using Jenkins, Helm, Kubernetes, and Terraform.

At Prifender I architected an AI platform for data discovery and record linkage. Because it is a start up I had to wear several other hats. I built ETL pipelines for Mongo, PostgreSQL, MySql, SqlServer, and Oracle. I prototyped GPU acceleration using Tensorflow in Python. When all the operations resources quit or were let go I filled in and managed customer deployments and resolved live issues in customer cloud environments. I also worked to gradually improve the build, deployment, logging, and monitoring processes of web services using Git, Jenkins, Ansible, ELK Stack, and Prometheus on Docker Swarm.

At Cision I was a team lead which focused on Java Spring microservices and artificial intelligence solutions at scale. In my first year at Cision I onboarded the company to Googles Cloud Platform using BigQuery, pioneered and integrated ElasticSearch, used NLP text classification and sentiment analysis libraries, integrated several Netflix OSS technologies, stood up a cluster using AWS EC2 + ECS, and led the transition from a manually curated VM+tomcat based deployment/hosting environment to Docker container orchestration. I'm particularly proud of pioneering and documenting developer conventions for Docker, AWS, and Kubernetes and driving their adoption.

During my years at Microsoft I built, managed, refactored, migrated to cloud, or sunset seven data pipelines related to Windows crash and usage data. I developed and maintained the entire stack of technologies required to acquire data, extract insights, and bring actionable conclusions to users within Microsoft and its partners. Technologies used include big data collection using clients and XML, .Net networking and encryption services to securely transport data, scale-out storage, management, and archival of Microsoft SQL and Cosmos data warehouses, SQL data marts, data reporting using .Net and SSRS, and data analysis using Sho in IronPython.

As a research assistant I have been the primary and co-author on several journal and workshop publications, book chapters, and NSF grants in conjunction with the IMPACT lab under Sandeep Gupta. My area of expertise centers on applying Artificial Intelligence solutions to Cyber-Physical-System domains.

At Northrop Grumman my work centered on designing real-time cognitive architectures and planning algorithms for coordinating fleets of unmanned aerial vehicles.