Verified Expert in Engineering
Shelley is a full-stack engineer with 15 years experience building software in a variety of industries. She is comfortable starting with rough requirements and working with stakeholders to turn an idea into a useful, appealing piece of software. Shelley writes clean, fast, well-documented, and well-tested code, provides realistic estimates, and works well with teammates. As a student, Shelley spent several years in an AI Ph.D. program and she maintains a strong interest in that field.
Git, IntelliJ, MacOS
The most amazing...
...work I've done recently was to build, from the ground up, a web application to benchmark my employer's threat detection capabilities against MITRE ATT&CK.
Senior Back-end Developer
Scratch Engine LLC
- Handled data engineering for an ML project learning to predict NFT prices.
- Worked with the ML engineer to locate data sources, like social media APIs, NFT, and crypto APIs, and to evaluate data quality.
- Built, deployed, and monitored the project's data pipeline (AWS Lambda functions writing to DynamoDB).
- Developed a detailed data inventory describing the data assets generated by the project.
- Architected and built the back-end services to support the project's web app in Node.js/TypeScript and Python.
- Architected and built the Use Case Manager, an internal-facing web application used by Kudelski's MDR team to benchmark their threat detection capabilities against the MITRE ATT&CK framework.
- Wrote the code for much of Secure Blueprint, a SaaS web application for CISOs to assess, manage and report on their company’s cybersecurity program.
- Built a POC API to provide medical cannabis product recommendations as part of a blockchain-related venture. Utilized a Drools-based expert system meant to bootstrap data collection for future machine learning.
- Served as an alpha customer for Kudelski's internal DevOps platform and contributed to its design.
- Introduced the idea of database versioning to my team and converted our database deployment process to use Flyway.
- Served as a careful and detailed code reviewer with a good track record of catching bugs in review.
Senior Software Developer
- Worked on greenfield development of Java web APIs to be used by all customer-facing applications at a major airline.
- Influenced the decision to use a JWT-based client-side cache, which greatly simplified the API while maintaining performance and statelessness.
- Trained junior developers through pair programming.
DEVEXI (Toptal Client)
- Wrote Java and SQL to execute retrospective longitudinal studies over a database of medical records and correlated environmental/socioeconomic data. Queries ran on AWS Redshift over billions of records.
- Sped up a central algorithm for matching cases to controls to run 50 times faster. Optimizations included re-representing a join table to be 10X smaller and reducing the size of the input by randomly sampling from the set of controls.
- Coded data perturbation and suppression algorithms to satisfy HIPAA de-identification requirements.
- Designed HTML emails for invoicing and built the service that generates and sends invoice emails.
Georgia Tech Research Institute
- Developed a new SPA-style web app for administering firewalls and specifying access policies for classified web-based services.
- Overhauled a complicated GWT-based application module for handling the workflow for requesting and approving personnel access to classified programs.
- Designed and implemented a stand-alone service to perform content inspection on uploaded files for use in a classified file-sharing system.
- Integrated new and existing applications into an OpenAM-based single-sign-on framework to support access-controlled user sessions based on authorizations returned by the SSO server.
- Profiled and diagnosed performance issues in our SSO service until it met performance requirements and doubled the requests/minute.
- Offered meaningful and polite feedback in hundreds of code reviews for teammates and student co-ops.
- Designed and implemented the next version of Veloxiti's core product, a C++-based cognitive engine, as part of a small development team.
- Developed a graphical knowledge editing tool for Veloxiti's cognitive engine, built as a set of Java plugins to the Protege ontology editor.
- Built a Java application to extract structured data from military operation orders and tactical chat using the GATE natural language processing toolkit.
- Implemented Java and Soap-based web services to communicate among Veloxiti's cognitive engine, a tool processing tactical mIRC chat, and the Army's Command Post of the Future.
- Performed background research, interviewed experts, and wrote the final report for an SBIR Phase I projected to build an expert system to assist in estimating Human Systems Integration costs in Naval acquisitions.
Use Case Manager (Kudelski Security)
I participated in the planning and design from the project's start, and I architected and built most of the app's services and front end. Kudelski was developing its own Kubernetes-based DevOps platform at the time, which allowed me to use a microservices architecture that included a collection of Java and Node.js services, variously attached to PostgreSQL, Elasticsearch, and Redis containers for data and fronted by a React/TypeScript UI.
Secure Blueprint (Kudelski Security)https://kudelskisecurity.com/services/secure-blueprint/
I joined the project immediately after the completion of the MVP and was the lead developer after that. I wrote the functional specifications and most of the code for many of the modules described in the linked website, specifically initiative planning, operational metrics, and board communication. My proudest project accomplishment was finding a relatively painless way to export slides designed in the browser by the user into a PowerPoint deck for board presentations.
Soar-RL (University of Michigan)https://www.sciencedirect.com/science/article/abs/pii/S1389041704000646
As part of my graduate school research, I built a version of the Soar cognitive architecture that incorporated a new reinforcement learning mechanism. This work was published in the journal Cognitive Systems Research.
Git, IntelliJ, Jira, Auth0, Gradle, GitLab CI/CD, Postman, Flyway, Docker Compose, Apache JMeter, Apache Maven
Agile Software Development, Scrum, Automated Testing, Unit Testing, Microservices Architecture, DevOps, Data Science
Web Development, Full-stack Development, Full-stack, Code Review, Back-end, Data Structures, Integration Testing, Multithreading, Algorithms, Machine Learning, Natural Language Processing (NLP), Artificial Intelligence (AI), Logback, Prometheus, OAuth, APIs, Non-fungible Tokens (NFT), Serverless, GPT, Generative Pre-trained Transformers (GPT)
Google Guava, JUnit, Spring, Mockito, Spring Boot, Hibernate, gRPC, Redux, Cypress
JDBC, Jackson, Node.js, React, REST APIs
MacOS, Linux, Kubernetes, Docker, Amazon Web Services (AWS), AWS Lambda
PostgreSQL, JSON, Redshift, Elasticsearch, MySQL, Redis, Amazon DynamoDB
Master of Science Degree in Computer Science
University of Michigan - Ann Arbor, MI, USA
Bachelor of Science Degree in Mathematics
University of Chicago - Chicago, IL, USA