In my role, I'm a machine learning engineer and research scientist solving complex problems in computer vision and multimodal learning in the domains of generative AI, 3D scene reconstruction, and biometric recognition.

• Developed an end-to-end system in Python using deep learning to extract and assign definitions in legal texts. This feature allows users to quickly access links to definitions when reading legal documents.• Leveraged the inherent structure of legal corpora to optimize algorithmic efficiency, reducing the overall system runtime by 90% .• Tested and prepared the system to be deployed onto the website of the Cornell Legal Information Institute. This institute aims to increase the… Show more • Developed an end-to-end system in Python using deep learning to extract and assign definitions in legal texts. This feature allows users to quickly access links to definitions when reading legal documents.• Leveraged the inherent structure of legal corpora to optimize algorithmic efficiency, reducing the overall system runtime by 90% .• Tested and prepared the system to be deployed onto the website of the Cornell Legal Information Institute. This institute aims to increase the accessibility and readability of legal publications in an effort to empower individuals without a legal background. Show less

• Graduate TA for the Computational Genetics and Genomics course at Cornell University. • Organized and taught biweekly lectures to a class of over 60 students, reviewing algorithms and methods relevant to computational biology.• Held weekly office hours to reinforce student understanding of the theory and algorithms presented during lectures. • Graded student assignments focused in both algorithm design and theoretical modeling and statistics.

• Led research efforts towards facial pose augmentation as part of a full-body recognition system.• Cultivated new methods for synthesizing photo-realistic faces using 3D GAN models and super-resolution techniques, leading to a 50% increase in recognizability across multiple views.• Made a key contribution in overcoming a domain gap between augmented samples and real-world data collections using both image-to-image style-transfer and GAN inversion techniques.

• Applied astronomical pattern recognition methods to identify whale sharks using their natural patterns of their skin under professor Chuck Stewart.• Designed a geometric-based approach in C++ that outperformed a deep learning model with 82% top-1 recognition accuracy. • This research was done as part of the WildMe project, which leverages machine learning for the benefit of wildlife conservation and sustainability efforts.

• Conceived a novel generative model framework for anomaly detection on attributed networks using normalizing flows and Graph Convolutional Neural Networks. Algorithms were implemented using PyTorch.• Curated custom datasets by designing a framework for generating and injecting artificial anomalies into existing networks. This was critical for generating positive and negative samples for model evaluation.• Improved upon state-of-the-art benchmark models by over 30% (F1 score) on road… Show more • Conceived a novel generative model framework for anomaly detection on attributed networks using normalizing flows and Graph Convolutional Neural Networks. Algorithms were implemented using PyTorch.• Curated custom datasets by designing a framework for generating and injecting artificial anomalies into existing networks. This was critical for generating positive and negative samples for model evaluation.• Improved upon state-of-the-art benchmark models by over 30% (F1 score) on road network and social media datasets.• Played a pivotal role in displaying the efficacy of our models against state-of–the-art approaches, leading to the funding of an IBM project applying our model to anomaly detection on knowledge graphs. Show less

Computer Science Tutor for Data Structures at Rensselaer Polytechnic Institute. I worked with students to help understand course materials and implement their designs on weekly assignments.