Benjamin Black Email & Phone Number

* Developed 3D z-stacked, independent channel YOLO based deep learning algorithm for fluorescence in situ hybridization (FISH) signal count differentiation, with significant ablation studies showing it was significantly superior to more generic methods. Also advised data labeling and validation efforts in the project, which was tricky due to the high level.

Remote

* Designed, implemented, and deployed a new high throughput queue-based ML inference pipeline 3x cheaper and more maintainable, by designing, getting buy-in from leadership, building an efficient pipelined implementation. Deployment occurred incrementally in a challenging regulated environment where all results must be identical before and after the.

Idaho Falls, Idaho, United States

* At Accendero, we make bespoke software for pharmacuticals to aid their small drug discovery processes and assist with regulatory compliance.* I lead a small team in a large effort to refactor a rigid, hardcoded full stack assay storage and reporting stack into a modular and flexible multi-assay system for accendero's largest client.* I developed a clean.

College Park, Maryland, United States

* Worked closely with team to design and implement PettingZoo, the defacto Multi-Agent environment library for Reinforcement Learning.* Was go-to resource on team for deep dives into external codebases, code review, debugging, resolving user issues, maintaining tests, and refactoring.* Drove the API design changes for PettingZoo such as the parallel API.

* Worked with Prof Mark Bedau and his group to develop visualization tool for understanding the behavior of running word2vec text embedding algorithm on MIDI music files* Independently developed contrastive learning method for making embeddings of audio files and visualization tool to understand the results (writeup linked).

* Spring semester: Intro to Computer Science TA: I graded quizzes and projects and tutored many students in drop in tutoring. * Fall Semester: Computability and Complexity TA: Graded extremely complex open ended questions with many possible answers and no rubric. Discussed with professor to resolve ambiguities.

Seattle

* AWS Network Automation helps automate setting up network devices in Amazon data-centers across the globe. * Fixed complex legacy code to allow automatic static management IP assignment to network devices, so that network engineers would not have to hand-allocate IPs. * Introduced complete diff testing for my changes, ensuring that IP-mismatches would not.

Portland, Oregon Area

* Full stack development, using Python, the Flask web framework, the SQLachemy ORM, PostgreSQL, HTML5, CSS3, and javascript.* Mentoring less experienced teammates in programming concepts and code details.* App graphic design.

Master'S Degree, Computer Science

Selected classes: * Advanced deep learning theory * Computational linguistics I * High Performance Computing

Bachelor’S Degree, Mathematics And Computer Science, Junior

Activities and Societies: ACM ICPC practice group, Orchestra, Running Club.Selected classes * Probability (upper level) * Advanced.