DevOps Software Company with industry standards like Terraform and Vault
- Engineered essential release pipeline onboarding infrastructure using Go, Github Actions, Docker Hub, Datadog, Vault, and Terraform, reducing onboarding time for 120+ future repos by 95%
- Authored a RFC that streamlined the architecture and onboarding process for 20+ engineering teams.
- Integrated unit and regression tests with Go mocks, achieving 90%+ coverage
- Resolved a critical CI/CD issue in Terraform releases, using Bash to auto-generate Github Issues related to Iron Bank.
World’s leading producer of educational services & software
- Developed an AI-driven exam booking chatbot for the Pearson VUE homepage, under Agile methodology, with projections to save $50k/year on customer support.
- Designed and demonstrated the implementation with a reactive homepage UI via Django, Javascript, HTML, and CSS, leading to C-Suite approval for project scaling and reducing future development time by 20%.
- Deployed a service to transfer chatbot conversation to Salesforce for further assistance seen by 40k+ site visitors.
Evolution and Ecology Lab (King Lab)
- Engineered scalable data processing pipelines in Python to enrich ecological and weather Excel datasets via external APIs, increasing data collection speed by 90%.
- Utilized Scikit Learn and PyTorch to train linear regression, KNN, and Neural Networks to predict virus and microbe concentrations across 100+ sample sites.
Omnistudy | Top 100 @ Google Solutions Challenge
AI education startup & Google Solutions Challenge Semi-Finals Winner. I served as the lead backend engineer
- Backend Software Lead for 3 other engineers, delegated tasks through Jira, accelerated project milestone delivery 70%
- Designed a MySQL schema and implemented CRUD in TypeScript & Node.js for user and document tables.
- Built a CI/CD pipeline leveraging GitHub Actions, resulting in a 50% reduction in deployment times to Google’s Firebase.
- Devised R.A.G. pipelines via GPT-4 LLM and Pinecone Vector Database achieving 40% higher accuracy than prompting.
plotsearcher.com
- Pioneered a plot-based book search engine, achieving a 45% increase in accuracy compared to traditional searches.
- Developed a ranking algorithm combining NLP with web-scraped Wikipedia plots, and scaled it using Locality Sensitive Hashing to add 100 query-to-book matches per day.
- Deployed the site on AWS (EC2) with a MySQL backend, built in JavaScript, Python, and Django with CSRF protection, email authentication, and password reset.
Gopher Robotics
Finished 23rd in our division, 10 places higher than last year
- Developed python scripts that utilize the Rospy library to devise algorithms for fully autonomous robot decision-making.
- Aided in the process of Redesigning Nvidia source code for optimal rough surface transversal via the addition of second esdf layer, augmentation of the costmap, and creation of a ROS wrapper.
