About

Devjit Chakravarti is an experienced data scientist with a passion for solving high-impact, challenging data problems using the right approach for the job. With over 16 years of relevant experience, he has a depth of knowledge across various domains, including machine learning (traditional and deep learning), telematics, signal processing, and sensor data analysis, experimentation, causal inference, product/growth analytics (e.g., retention, churn, funnels), forecasting and time-series analysis, data engineering (e.g., ETL, schema design, operations), visualization, business strategy, presentation, and storytelling, and large-scale (distributed) model training and development. Devjit is very comfortable working with both Python and SQL environments and has substantial experience with a wide array of technologies, including Tensorflow/Keras, Spark, Kubernetes, SKLearn, Airflow, Presto, Hive, Redshift, BigQuery, Tableau, Mode, D3, R, and more. He is a multi-faceted data scientist who strives to be an expert at anything he does. Currently working as a Machine Learning Data Scientist at DoorDash, Devjit has led the development of a simulator for the assignments system from scratch, including dasher behavior models, infrastructure, and workflow management. He has also developed a novel deep learning ML approach to food prep prediction, reducing RMSE by 20%, and authored a suite of internal ML development tools for profiling data quality, simplifying model training, and interfacing with internal data systems. In the past, Devjit has worked as a Lead Research Scientist at Lyft, where he led a team extracting driving risk signals from mobile sensor data (aka telematics). He has also developed and deployed multiple machine learning models, including Keras/Tensorflow Sequential Neural Networks, GBDTs & KNNs for improving travel time and traffic estimation, built and analyzed experiments for understanding causal relationships between retention & customer engagement, and developed and deployed new dynamic pricing models. Devjit has also worked as a Senior Data Scientist, Product Analytics at Facebook, where he supported Instagram & Parse with analytics and product insights. He was the lead analyst on Instagram Growth, helped develop international growth strategy through quantitative and qualitative research, informed NUX design through cluster visualization of public content following, performed lifecycle/funnel analysis, clustering, experiment analysis, forecasting, competitive analysis, and provided cross-functional support for Product, Engineering, Marketing, Sales. Devjit holds a BS in Aerospace Engineering from the