(𝐀𝐂𝐐𝐔𝐈𝐑𝐄𝐃: By Alphasense in 2022)Smart Synonyms for User Querieso Developed real-time synonym generation of user queries to enhance the search experienceo Created a text corpus of 116M words using financial docs of more than 71k companieso Trained embeddings from scratch on word corpus using AWS Sagemaker Skipgram modelo Generated Synonyms based on Nearest neighbours using Cosine similarity Trending Terms of companies for every financial quartero Terms are calculated based on a Count score in document normalized by min-max scalero Count score is then adjusted on a time relative scale based on its historical frequency o Overall term sentiment score is also calculated to highlight the sentiment of that term

Built from scratch, a dynamic pricing model based on competition property pricing: o Developed star rating prediction model on OTA data through Random Forest Classifier o An amenity matching score based on Nearest Neighbours to capture similarity in amenities o Similarity based on property description using word vectorization though GloVe Embeddings o Led to accuracy of amenity matching close to 87% and ultimately a 5% Revenue incrementBehavioural clustering of properties based on their intrinsic attributes o Segmentation helped in treating similar properties as a group in pricing and revenue initiatives o Used attributes of property like ADR, Occupancy, Amenities, Age to cluster in 5 segments o K-Means and spectral clustering used to achieve silhouette score of 0.73Awarded the coveted CEO Performance Recognition award for driving pricing strategies during COVID

(𝐀𝐂𝐐𝐔𝐈𝐑𝐄𝐃: By $11bn TVS-backed Carcrew)'ClickGarage' is an on-demand car and bike services platform. ClickGarage is building a network of standardized doorstep service providers and workshops (called "ClickGarage Workshops") with an aim to make quality car/bike maintenance services affordable, convenient and transparent. As a Data Scientist at ClickGarage, I was involved with following tasks• Developed end-to-end Automated Machine Learning(Auto ML) model • Built predictive analytical models to extract the patterns and behaviour insights• Database visualization and analysis to improve the targeting of customers through marketing• Critically assessed data and built a statistical model for the pricing of different services

Designed Eco-friendly concrete using the industrial wastes of Granite and Marble PowderAnalyzed the compression and sheer strength data of concrete blocks to achieve optimum accuracyUsed tools like Python, MATLAB, STAAD, Tableau for Data Preparation and Visualisation