Kishan G

Experienced professional in process improvement, public relations, and relationship management. Committed to driving results and adding value to your organization. Reliable, hardworking, and collaborative. Precocious Data Scientist Trainee ready to accept increasingly complex challenges associated with maintaining and exploiting growing data stores. Driven to expand experience through hands-on training and guided participation in effective data management tasks. Ready to immediately contribute beneficial input to employers.

Computational Data Science Training -Apr 2023

• Developed a predictive model: Created a machine learning model using Python/R to predict assessed property values based on various features such as property size, location, amenities, and historical data.
• Data preprocessing and feature engineering: Performed thorough data cleaning, handled missing values, and conducted feature engineering techniques to extract meaningful insights and optimize the predictive model's performance.
• Implemented regression algorithms: Utilized advanced regression algorithms, such as linear regression, decision trees, random forests, or gradient boosting, to train and validate the predictive model for accurate property value estimation.
• Optimized model performance: Employed techniques such as cross-validation, hyperparameter tuning, and ensemble learning to enhance the model's performance metrics, such as mean absolute error (MAE), root mean squared error (RMSE), or R-squared value.
• Presented insights and results: Prepared visualizations and summary reports to effectively communicate the findings, model performance, and actionable insights to stakeholders, demonstrating the potential impact on property tax assessment accuracy and efficiency.

• Built a fraud detection model: Developed a robust machine learning model using Python/R to identify fraudulent auto insurance claims by analyzing various data points such as claim details, policy information, customer history, and external factors.
• Data preprocessing and feature engineering: Conducted extensive data preprocessing techniques including data cleaning, handling missing values, outlier detection, and feature engineering to extract relevant features and enhance the model's ability to detect fraudulent patterns accurately.
• Implemented classification algorithms: Utilized advanced classification algorithms such as logistic regression, decision trees, random forests, or gradient boosting to train the predictive model and classify insurance claims into fraudulent or non-fraudulent categories.
• Feature importance and pattern extraction: Utilized feature importance techniques like permutation importance, SHAP values, or information gain to identify the key variables that contribute most to fraudulent claims. Extracted meaningful patterns and insights from the data to better understand the fraud patterns in auto insurance claims.
• Evaluated model performance: Employed appropriate evaluation metrics such as accuracy, precision, recall, F1-score, or area under the ROC curve (AUC-ROC) to assess the performance of the fraud detection model. Achieved high accuracy in identifying fraudulent claims, thus enabling the insurance company to mitigate financial losses.

Application of Node Frequency routing to Detection of Packet loss in Wireless Ad-Hoc Networks (06/2022 - 11/2022)

• To design and implement a sequence in order to identify the malicious node, which is the cause for the packet loss in Wireless Ad Hoc Network.

KNN Classification Algorithm under INSOFE (02/2023 - 04/2023) 

• During my internship at , I worked on the development of the K-Nearest Neighbor Classification algorithm for the No-Code-AI project.

Machine Learning with Python (01/12/2021 - 31/01/2022)

• Successfully completed the Machine Learning with Python live Projects program from Zebo.ai in association with Verzeo, gaining hands-on experience in applying machine learning algorithms to real-world projects.