About

Vaishak Kumar is a Machine Learning, Data Science, and Econometrics practitioner with a passion for building thinking machines. With a background in Economics and Robotics, he is well-equipped to implement cutting-edge systems from just reading academic papers or engineering blogs. His interests lie in natural language processing, computer vision, statistics, and experiment design. Vaishak is currently working as a Data Scientist/Machine Learning Engineer at Stimscience, where he is finding optimal brain stimulation protocols to increase sleep quality, classifying EEG sequences into Sleep Stages, and conducting causal inference and study design. Prior to this, he worked as a Software Engineer at Yelp, where he laid out a vision roadmap for improving image selection model using neural networks and further using its output as a feature in ad candidate selection models, assessed correlation of advertiser tenure time series data with system warm-up data, and proposed strategies to reduce churn. He also implemented product features using ElasticSearch, conducted data analysis on A/B experiments to increase ad quality and relevance using Redshift and Spark, and proposed and implemented usage of query understanding to improve user ad experience. Before Yelp, Vaishak worked as a Computer Vision Research Intern at SRI International, where he was responsible for Occlusion Reasoning for Augmented Reality using Instance and Semantic Segmentation, contributed to Semantic Navigation tasks with different modalities (Images and LiDAR), and adapted deep learning models for real-time systems and systems integration with C++ and Python. Vaishak holds an MSE in Robotics from the University of Pennsylvania, an MSc in Economics from the London School of Economics and Political Science (LSE), a Nanodegree in Artificial Intelligence from Udacity, and a Bachelor of Arts (B.A.) in Economics and Political Science from the University of Pennsylvania. He is proficient in programming languages like Python, Java, and JavaScript, and data science libraries like Pandas, NumPy, and Scikit-Learn. He also has experience working with big data libraries like RedShift, ElasticSearch, and Spark, computer vision libraries like OpenCV and TorchVision, and natural language processing libraries like Gensim and spaCy. Vaishak has worked across design, implementation, and testing of backend and machine learning systems and has a relevant experience of 3.62 years.