About

Jainesh Doshi is a Senior Applied ML Engineer at JP Morgan with a focus on Natural Language Processing (NLP) and time series forecasting. He has a Master of Science in Computer Science from the Georgia Institute of Technology and a Master of Technology and Bachelor of Technology in Electrical and Computer Engineering from the Indian Institute of Technology, Bombay. Jainesh has a wealth of experience in developing deep neural network models for predicting equity returns and has handled transactions worth $15 million USD every day. He has also built a firm-wide text summarization tool using Longformer-based models for event-centric clustered documents. Jainesh is currently working on developing (abstractive + extractive) summarizers for event-centric clustered news articles utilizing Longformer Encoder Decoders. Jainesh has also worked as an Applied ML Engineer at JP Morgan, where he was responsible for developing LSTM-based models to predict returns for 1100 equities across markets handling $15 million USD a day. He has also built a real-world analogous stock market simulation environment with Limit Order Book & trade queues and trained model-based RL agents to optimize trade breakdown and execution. Jainesh has devised model regulatory approval strategy & formulated metrics to analyze model performance & signal persistence. He has also investigated multi-step and multi-horizon prediction for non-stationary time series using encoder-decoder models. Jainesh has also worked as a Graduate Teaching Assistant at the College of Computing at Georgia Tech, where he taught Game AI under Prof. Mark Riedl. Jainesh's technical skills include ML, NLP, Sagemaker, Research Scientist, and AI. He has a relevant experience of 4.20 years. Jainesh is a mentor at the Silicon Valley JPMC SEP Program, where he hosts showcase events and fireplace chats with leaders. He has also built a daily report generator summarizing worldwide critical cybersecurity-related incidents to senior executives. Jainesh has deployed a transcript summarizer on Sagemaker for on-demand execution optimizing resource usage.