Analytical mindset, with broad expertise in extracting valuable information from data. Enjoy open-ended problems, and the use of emerging technologies for solving challenging cases in a wide range of industries. Having research and consulting experience, in combination with strong analytical skills, enables me to approach complex problems through mathematical models and computational algorithms. Significant experience with AI/Machine Learning and data analysis using e.g. Python. Proficient with machine learning tools such as Scikit-learn, TensorFlow and Keras. Currently leveraging H2O.ai as well to solve problems having highly imbalanced datasets. Looking to deliver insights and provide action-oriented solutions to complex business problems at my next data science role.
Aditya Gautam is a highly skilled Data Scientist with a keen analytical mindset and expertise in extracting valuable information from data. He enjoys working on open-ended problems and using emerging technologies to solve challenging cases across a wide range of industries. With research a
Years of Experience
university of pittsburgh, mahe
indusind bank, iffco tokio general insurance, r systems, university of pittsburgh, lt
Reach out to Aditya Gautam via Email, InMail and SMS drip
by installing Chrome extension
Aditya's contact details
2022 - Present
Data Scientist (AVP)
2021 - 2022
iffco tokio general insurance
Data Scientist (Asst. Manager - official title)
• Fraud Detection. Using tree based and gradient boosting classification methods with synthetic sample generation for optimal class balance. • Customer Segmentation. Using K-Means algorithm to segregate customers in groups for optimal marketing strategy and improving customer experience. • Renewal Propensity. Identifying customers based on the exhaustive profile which will renew their car insurance policy. Using boosted trees the renewal rate increased by 5-7% on avg. as compared to last year. • Using Image classification and detection models developed a tool for automated claims processing for private car product.
2019 - 2021
• Developed algorithms for predictive maintenance for the machines used in manufacturing process which led to 15% increase in manufacturing efficiency. • Made Recommendation System for an European e-commerce client. Based on the buying behavior of the customers, generated the likelihood scores for combinations of products being bought together. • Customer Segmentation. To differentiate customers into groups based on their spending patterns, ordering frequency etc. to optimize the marketing strategies. Which led 5% increase in overall sales within 2 months of implementing the solution. • Anomaly Detection for a Telecom Network provider. This solution helped the client in identifying the patterns for the unusual spikes in network usage. • Identified key features on the calls, for a big ISP, that affect the sale of a product using binary classification with custom score optimization methods with data gathering from MicroStrategy and Hadoop data lake. • Uplift modelling for to find out the effectiveness of agent training for call behaviors and product offerings. • Developed deep learning model for identifying the tabular structures and other key value pairs from invoice type documents. Made it a language agnostic solution as it utilizes CNN.
2017 - 2019
university of pittsburgh
Data Science Student Consultant
Projects: • Drug Interaction: Predicting drug-drug interactions using the drug substructure, target, side effect, off label side effect, delivery methods and more features. The dataset contained over 150 features and 2 major interaction classes (PK) and (PD) and one was categorized as unknown. Using XGboost classifier, Random Forest, logistic regression, neural network and feature reduction methods(LASSO, PCA) a comparison study against labor intensive lab methods was conducted. Jan.2018-Apr.2019 • Analysis of Athletes: Analyzed the performances of amateur athletes in the Ironman championships from 2011 to 2015. Found out which event of the race was most influential in determining the outcome using Regression and after baselining the performance metric an exponential curve was fitted to determine the deterioration constant for the performance and the age where performance equals 0 using Scipy optimize. Aug-Dec 2018 • Taxi Analytics: Finding out the factors affecting the fare amount for yellow cabs in NYC and suggesting an optimized strategy for drivers to improve their profits, using statistical Machine Learning (linear and non-linear regression, K-NN, neural nets, PCA, SVM) algorithms on a 1 million+ rows dataset. Jul-Dec 2018 • OR Inventory: Analyzed data for current stock and inventory levels for endocrinology department and determine opportunities to reduce and/or optimize purchases and stocking levels. Developed plans for inventory storage and improving the processes for replenishment and for inventory monitoring in collaboration with Premier Inc. and Stryker Sustainability at Ruby Hospital WVUHS. Aug-Dec 2018 • Research: Assisted in the research for Methods for Surgery Duration Estimation. Developing Prediction Bounds using the variety and distinctness of historical Surgical Data. Classified the complexity of the procedure based on judgements from the experts and test it against one year of actual cases. (UPMC) Oct.2017-Jun.2018
2014 - 2014
Project Intern • Learned the various aspects of project management at a major Delhi Metro Rail Corporation site. • Increased the efficiency by 15% in Procuring and logistics department using lean methods. • Under the guidance of Site engineer reporting and maintaining quality control.
Artificial Intelligence (AI)
Business Intelligence (BI)
Convolutional Neural Networks (CNN)
Hypertext Transfer Protocol (HTTP)
Machine Learning (ML)
Support Vector Machine (SVM)
2017 - 2019
university of pittsburgh
Master of Science - MS
2011 - 2015
Bachelor of Technology - BTech