Reston, VA
john@nevin.dev
(978) 973-6262
Education
Work Experience
-
Member of the Trusted Postgres Architect (TPA) product team.
-
Led development of an automatic file checksum validation feature to pinpoint troubleshooting of deployed clusters by comparing Trusted Postgres Architect (TPA) installations in customer environments against the released package.
-
Led migration of Github Actions CI/CD workflows as part of the open-sourcing effort for TPA, increasing visibility and adoption.
-
Developed a semantic similarity search tool to improve integration testing by comparing vector embeddings of customer’s deployed clusters against our test clusters.
-
Implemented ability for TPA to deploy a novel and widely-adopted distributed Postgres cluster architectural pattern for active-active data replication and failover management.
-
Led development of a TPA feature for upgrading software installed on deployed clusters with minimal to no downtime in availability, backup or failover management.
-
Developed Python applications to interface with third-party REST APIs to retrieve millions of events daily from many different sources and tools for forwarding, indexing and searching in Splunk for integration with Kibana and Elasticsearch.
-
Incorporated robust logging into Python applications and wrote statistical searches for monitoring performance and data quality.
-
Normalized, filtered and enriched terabytes of heterogeneous data with fast, consistent and maintainable searches in Splunk.
-
Worked to maintain 80% coverage of all data retrieved from disparate sources against the expected state and reduced downtime for reporting data downstream.
-
Fully automated the deployment of Splunk app releases with a Python program; handles backing up, configuration and deployment with a user-friendly interactive CLI wizard.
Technical Skills
Personal Projects
- A full stack Phoenix application to build lifting workouts track bar-paths using AI. Utilizes Elixir channels and websockets to enable real-time collaboration for program creation. Workout days are automatically scheduled on a calendar and exercise completion is tracked for analysis.
- Videos of lifts are uploaded, enqueued, analyzed and processed using S3, RabbitMQ, AWS Lambda and an OpenCV Flask API respectively. The job status is shown to the user in real-time and a URL with the analyzed video is provided upon completion.
- Infrastructure is deployed as code using Docker Compose, OpenTofu, and Tailscale for fast, secure and consistent deployment across different environments for development, testing and production.
Project Repository
A Progressive Web Application for reporting near-miss incidents between pedestrians/cyclists and motorists. Automatically records location of incident and allows users to easily provide details in dynamic forms. Thousands of map features can be displayed due to GPU-accelerated vector tile rendering. Web form questionnaires are built dynamically by parsing JSON. All data is stored locally in the browser IndexedDB for full-offline capability and user-ownership of their own data and sharing decisions
Awards and Achievements
Authored ‘Representing Graphs in PostgreSQL with SQL/PGQ’ which became the most-viewed post of the month with over 7,500 hits. Gained traction and generated discussion on both YCombinator Hacker News and lobste.rs boards.
Received $4,000 in funding from the university to further develop my hackathon-winning project into a
minimum viable product in an eight week accelerator program.
Won a 24-hour Hackathon for creating a smartphone application for assisting individuals with memory loss that addressed the hackathon theme of “safety”.