Vineeth Rakesh Mohan Email & Phone Number

San Jose, Ca, Us

San Jose, Ca, Us

Foster City, California, Us

Foster City, California, Us

AI Modeling Leadership: Led AI initiatives in fraud detection and transaction prediction, enhancing team innovation and efficiency. Created and executed deep learning models for heterogeneous structured data, significantly outperforming previous models. Cross-functional Collaboration: Collaborated closely with product teams to collect requirements and constraints, effectively managing resources and latency considerations within a dynamic environment. Documentation and Collaboration Enhancement: Developed extensive documentation and a centralized code base, boosting productivity and facilitating knowledge exchange between research and product teams. Generative AI Advancement: Directed the creation of a novel conversational AI model for large-scale transaction data, marking a significant business impact and raising project visibility within the organization.

Wilmington, De, Us

Patent Search and Retrieval System- Improved the ranking performance of the existing system by up to 20% by incorporatingcontextualized word embeddings.- Proposed and deployed Siamese Transfomer (BERT) model for comparing texts betweenlong documents and providing interpretable reasoning behind their (dis)similarity. Fileda patent for this algorithm.Few-Shot Learning in Neural Networks- Minimized the cost of data acquisition using active learning for image classificationand instance segmentation tasks. This work was published at CVPR 2021.- Reduced the training time of neural networks using Bayesian uncertainty techniques

Los Angeles, Us

Behavior modeling for Sling OTT media devices- Developed feature engineering pipeline using Spark MLlib that aggregates logs ofdevices and apps to create user attributes.- Improved existing app recommendation system by 10% by incorporating temporalsignals in collaborative filtering.- Proposed and deployed a scalable churn prediction model that predicts the drop-outrate of users with an accuracy of 82%. This work was published at RecSys 2021.

Tempe, Az, Us

Deep Learning Models for User Personalization and Information Retrieval: Focused onleveraging generative deep neural networks such as Variational Auto Encoders (VAE) for tasks such as link prediction, content summarization and feature learning. Proposed Linked VAE model for recommending items that can serve as supplements or substitutes to an item purchased by a user. Evaluated the model on large scale E-commerce dataset from Amazon.

Los Angeles, Us

Implementation of Information Retrieval System in Large Business Setting: Focused on developing latent aspect models for information retrieval systems. Proposed sparse topic models for retrieving fine-grained sentiments and aspects from item, movie and restaurant reviews from domains such as Amazon, IMDB and Yelp.

Philadelphia, Pa, Us

My internship primarily focused on developing scalable personalized recommen- dation models for X1 Xfinity System. During my internship, I worked on two key projects: my first project was to optimize for cache storage by clustering consumers of video on demand (VOD) programs. This task was performed by infusing scalable latent factorization techniques using Apache’s spark framework into Xfinit’s state-of-the-art recommendation model to provide personalized program suggestions on a cluster-level. In my second project, I worked on the problem of automated profile detection by developing hybrid collaborative and content based recommendation framework.

Detroit, Mi, Us

Developed recommendation and predictive models on heterogeneous data sources such as microblogging data from Twitter, location-based data from Foursquare, and crowdfunding data from Kickstarter. Proposed and implemented several variations of topic models and other machine learning models, which include ensemble, regression, and clustering techniques for applications such as recommending projects to crowdfunding communities, suggesting point of interests (POIs) to travelers, and capturing fine-grained topical summaries from Twitter.

Detroit, Mi, Us

I served as teaching assistant to the 5 level graduate course on datamining called “IntelligentSystems : Algorithms and Tools”, and the advanced 7 level course on “Data Mining :Algorithms and Applications”. During this period, I was responsible for teaching variousconcepts on text mining, and presenting demos on data mining tools like Rapid Miner, andOrange datamining software.

Detroit, Mi, Us

Worked as server support staff for Computing and Information Technology (C&IT) de-partment of Wayne State. I was responsible for maintaining and trouble shooting variouscomputers and server systems within the university’s main campus. My work also involvedwriting powershell scripts for automating server side installations.

Phd, Computer Engineering

Masters, Computer Science