Revitalized Ad Insertion Engine for HLS Streams: Rescued a non-operational ad insertion engine with zero documentation or institutional knowledge, restoring it to full functionality and scalability. Achieved monetization levels of approximately $8 million USD per month, supporting Adult Swim’s streaming services with a highly reliable ad experience. Including managing DRM enabled HLS streams in fairplay, widevine.Automated HLS Variant Registration: Identified and automated critical, previously manual HLS variant registration steps for an external service, significantly reducing overhead. Mentored junior engineers to automate dynamic HLS file generation, enhancing process efficiency, accuracy, and speed in response to performance demands.Streamlined Dynamic Ad Insertion Processes Using Golang: Spearheaded modernization of ad insertion workflows in HLS streams, leveraging Golang to design a simplified, robust, and maintainable system. Enabled rapid monetizable outcomes by creating a more agile, responsive ad insertion platform that significantly improved usability and reduced complexity.

Joined Neural Payments as a private engineering consultant to consult their ongoing engineering roadmap and help to automate and streamline customer onboarding both in and outside of the IDE. Constructed untethered, asynchronous, idempotent Tokio based Kafka consumer, Kafka producer, orchestrator, and all associated primitive traits for custom idempotent disbursement messaging via protobuf including custom serialization implementation.Refactored client code for reduced build time and optimized rust binaries. Lead and engaged in conversations related to technology roadmap and technology adoption for ongoing feature development.

Cross functional engineer focused on KTLO and coordinating across multiple teams for automated device enrollment features.Led a cross-functional initiative to modernize firmware update processes, resulting in $4.5 billion USD in IT savings by transitioning to untethered, idempotent, and asynchronous multi-device updates.Collaborated with leadership to establish standards for automated customer data ingestion, eliminating the need for extensive manual input.Utilized existing data engineering workflows to reduce expected work for workflow automation for device engineering by 90% or more, utilizing GCP object storage notifications to kick off container jobs managed by an apache airflow workflow. Personally embarked on a unique project to reverse engineer an elevator, showcasing a diverse skill set and a passion for hands-on problem-solving.Contributed to fundamental changes in business operations, demonstrating a proactive approach to problem-solving and innovation.

Fully designed internal golang plugin, typescript plug-in data service, and Amazon cloud services such as S3, serverless, and cloud formation to route protect and enable internal UI use case testing with real data on demand.Altered CI/CD Pipeline to coordinate installation of shared plugins for serverless framework across every Sevco Security Deployment, reducing redundant code shared across repository. Altered Translation table to better handle localization features and push unified language in React Frontend.Designed and developed AWS ingress service for JWT authorization, and dynamic route resolution policies for all AWS services at Sevco Security

Refactored legacy Django back-end & React front-end, improving process for checking collaborators on docker hub accounts.Removed duplicate sources of truth in UI and Golang Back-end for core account and organization information including private repositories for accounts and other information related to billing status with Docker Hub Subscriptions.Coordinated with staff level engineers, management, and business operations for plans to replace legacy billing platform with modern stripe integration.Directed and produced technical walkthrough videos for other engineers and business operations to understand various aspects of integrating stripe APIs.Updated event notification system to better encompass events on the Docker Hub platform for an enhanced user experience.Spun up the entire stack for a new service running on kubernetes, via helm, and integrated it with Jenkins CI/CD and third party observability tools.

Fixed a production race condition affecting 1% of revenue.Implemented endpoint for custom transcode only feature with custom ffmpeg options, allowing for specific ffmpeg filtering options for audio and video. Worked directly with technical lead to produce architectural design for golang static video packaging service, utilizing rabbitMQ, Prometheus integration, and Kubernetes to achieve maximum throughput.Integrated Prometheus metrics, such as job duration, common failures, etc. to support site reliability for supporting managed k8s instances.Integrated Kibana metrics for specific common failures in video transcoding jobs, allowing for searchable, detailed information for the site reliability team.Designed a slow spin down feature for rabbitmq powered service utilizing kubernetes signals to slow intake of jobs dynamically while still allowing maximum throughput to service during scale down.Implemented multi stage transcode plan executor in a distributed context to execute multi-threaded and multi-staged video transcodes with a distributed locking mechanism to accommodate more advanced features for video packaging with agnostic packaging services such as the one mentioned above.

Utilized golang to Author serverless software run as AWS lamda functions to forward various authentication requests fromOkta, Auth0, Google Oauth, etc. to a unified interal Oauth2 API.Utilized Jenkins CLI, Gotest, and other frameworks to achieve 100% unit test coverage and sufficient Integration testcoverage for all delivered production code in the SDLC (Software Development Life Cycle)Changes resulted in the ability to begin deprecation of 16 Legacy Auth API endpoints and reduce developer time spent maintaining logins.Implemented unified forgot password API endpoint using transformer architecture to reset passwords for every Oauthprovider

Designed, planned, and assisted in the creation of a rich debugging feature-set to profile for the Disney+ client experience on the next generation of consumer hardware. (PS5, Xbox, x86_64 & Arm televisions)Utilized memory management, client access patterns, and access diagrams to propagate information in a secure and safe manner, to expose and assist developers in finding hitches in one of the most complex and vibrant codebases in the world today; The Disney+ Project.Succeeded in reducing exposed data between debug specific code and the base client, preventing situations where arbitrary information exchanges could occur by instead utilizing features of the Rust language's macros to create predictable, well-defined behavior in a debugging environment.Additionally, these changes ensured that the debug support provided did not expose information in the platform in a way that either breached access patterns established for the rendering SDK or client engine, or could otherwise be substituted with information from a source that made sense in terms of the Rust ownership model.Obtained valuable mentorship and experience in planning large features and system design, producing diagrams and access pattern maps to facilitate clean updates and update senior engineering staff on progress, and to obtain meaningful feedback across changes with well documented approaches.Changes will result in significantly less need to re-compile, saving large swaths of developer time when considering the compile times at the scope of the Disney+ project.

Assisted in the maintenance and support of proprietary software systems at the Rawlings Group LLC. As a member of the Audit App-Development team I provided developer time to mitigate the burden of legacy, gather information for requirements, and modified processes to better adhere to company standards at the largest Insurance Subrogation firm in the United States.-Updated Image processing systems that were experiencing issues at scale due to bottlenecking-Code changes allowed for radical upscaling of support for two clients-Changes allowed for newer better supported hardware to support a 64 bit version of the service group-Significantly improved the speed and performance queries to the image database using clever indexing methods

Developed and researched novel methods of Pharmicovigilance in the drug space, utilizing Artificial Intelligence to assess risk of taking the drug Tamoxifen in elderly patients.Results of the continued study showed with peer review and reproducible results that co-morbidity and age can be a reliable indicator for adverse drug events especially compared to a Doctor’s intuition or opinion.Continued study necessary but results showed a 20% increase in accuracy using MDR based prediction schemes as opposed to any single group of predictions.