Sasidhar Sannadi

Developed recommendation model to recommend the product to users using LightFM algorithm, which combines collaborative and content-based filtering methods to tackle cold start problems. The model recommends new products to stores and stores to new users.

Worked on the entire project cycle, defining requirements and implementing solutions for the Cricket Portal using SharePoint 2010.

Worked on the entire project cycle, defining requirements and implementing solutions for SharePoint 2013. Configured various SharePoint components (SSRS, Load balancer, Distributed cache) and developed the CH portal extranet.

Involved in extracting and preprocessing data from ArcGIS using Python APIs. Performed visualization and stored data using SQLite and ODBC connection to Qlik Sense. Developed a Maara chatbot using AWS Lex and Lambda, integrating it with Inwards app and storing data in AWS DynamoDB. Also scraped real estate data, visualized it using Matplotlib and Seaborn, and trained models to predict profitable business locations.

Performed clustering and EDA on patient records, segmenting patients (TB, OldTB, Asthma, Healthy). Trained models on clustered data for prediction. Generated Spectrograms from WAV files using librosa STFT and trained a CNN model to detect and classify coughs (TB vs Non-TB).

Conducted clustering on 8 million records, generating features using RFM Clustering to segment customers (High/Mid/Low value) and calculating LTV. Used Apiary algorithm for Market basket analysis. Also designed alternate credit scoring models using mobile data, performing EDA on 25k+ customers, and implementing text classification (Logistic regression) on SMS text to predict probability of default. Prepared customer sentiment analysis.