About
CTO and CRO here @ Hyperbound. We're a YCombinator startup building the next generation of AI GTM.
Experience
Ads Machine Learning Engineer
Meta · New York, New York, United States
Applying SoTA Semi-Supervised Learning techniques to the most critical models in the Ads Ranking Pipeline
Machine Learning Researcher
Carnegie Mellon University · Pittsburgh, Pennsylvania, United States
Built classifier for large scale ICD-10 code Ranking and Classification for ER admissions Modularized classification and metric architecture for faster development iteration
Ml/Nlp Researcher
Carnegie Mellon University · Greater Pittsburgh Region
Utilizing dependency parsing (NLP) to extract technical information from patents My latest CNN model identifies software patent claims with 93%+ accuracy Publishing research on various NLP and ML based classifiers this year
Machine Learning Intern
Implemented Meta-wide multimodal contrastive loss classifier based on recent research Engineered several features with 6-point gain for Creator Classification and Segmentation Developed self-supervised user embeddings to improve segmentation generalization
Machine Learning Intern
Trained ML models to capture complex relationship between hyperparameters and notification metrics Utilized candidate generation, ranking, and filtering to suggest optimal hyperparameters Achieved tentative metric wins in unwanted notification volume and comments
Software Development Engineering Intern
Developed transfer learning strategies to improve search results, widgets, and ads in more than 6 marketplaces Engineered machine learning features from combinations of search query data in several locales Productionized flexible cloud infrastructure for massive language model development
Software Engineering Intern
Lockheed Martin · San Francisco Bay Area
Implemented chaos engineering with Gremlin to improve security and stability of cloud deployments Integrated container automation and validation software in Docker-In-Docker sandbox on AWS GovCloud Led team using Agile methodologies to develop highly-available cloud infrastructure as code on AWS
Research Intern
Rensselaer Polytechnic Institute
Developed a feedback control for advanced lighting systems with Newton’s method Equalized lighting intensity and color regardless of changing external light sources Weighted color equality, cost savings, and aesthetic appeal in multivariate optimization
Education
Y Combinator
S23
Carnegie Mellon University
Master of Science - MS, Machine Learning
Carnegie Mellon University
Bachelors, Computer Science with Concentration in Machine Learning
Find leads like Atul
Origami finds your ideal customers across 15+ data sources with a single prompt.
Try Origami Free