About

Raghuram Nagireddy is a highly experienced Lead Data Scientist with over 8 years of experience in building and leading Machine Learning systems involving data collection, processing, feature engineering, predictive modeling, and production deployment at scale. He is currently working as a Senior Manager, Data Science at Tiger Analytics. Raghuram has delivered measurable outcomes at leading organizations across diverse industries such as Automotive, Finance, Advertising, and E-Commerce. Raghuram has strong academic foundations gained at leading universities such as Columbia (NY, USA) and IIT-Madras (TN, India). He has extensive hands-on experience with a wide range of Data Science tools and libraries in Python and R. He has a TechStack that includes Data Scientist, ML, Research Scientist, Big Data, Python, AWS, Keras, Scikit, Tensorflow, Azure, Supervised Learning, SVM, Java, Deep Learning, Search, and Computer Vision. In his previous role at Fujitsu, Raghuram built large scale solutions to complex optimization problems (of NP-Hard nature) in vehicular networks (V2V), multi-modal transport, and computational fluid dynamics (CFD) for race car design using Fujitsu’s Quantum Inspired Technology, Digital Annealer. He also built a novel personalized recommender system for high-dimensional data for golf equipment and a predictive model for tomato plant yield as a function of various sensor measurements such as CO2 level, humidity content, plant weight, etc. He designed large scale big-data architectures on Azure, AWS, and Cloudera. He extensively worked with Python, R, Java (including a wide range of Machine Learning libraries such as Scikit-learn, Keras, Tensorflow, NLTK, etc.) and visualization tools such as Tableau, Qlikview, and Power BI. He developed all solutions with a high degree of collaboration with the stakeholders, driving a direct business impact. Prior to Fujitsu, Raghuram worked at Goldman Sachs as a Data Scientist. There he led the Batch Behavioral Model project to predict failures of a large number of batch jobs (~1 million) which have a complex implicit dependency structure and create alerts if a critical milestone is predicted to be breached. He designed a large-scale architecture for data storage, cleansing, transformation, standardization, etc. as input preparation for the predictive models. He leveraged a plethora of tools and techniques in Machine Learning such as PCA, Auto-encoders, Logistic Regression