Aspiring Data Scientist. 3 years of experience in AI/ML Engineering, specializing in Finance and Banking. Proficient in in developing machine learning models, leveraging strong programming skills and expertise in AI/ML frameworks. Hands-on experience with Python, TensorFlow, Keras, and Scikit-learn for model development. Strong understanding of data pre-processing, statistics, feature engineering, and model evaluation.
-Worked on a predictive model which analyses transaction data and detect potential policy violations. such as personal purchases or unauthorized expenses.
- Fine-tuned the model by trying out various algorithms and changing the hyperparameters for different algorithms.
- Collaborated with stakeholders to gather feedback, iterate on models, and continuously improve predictive accuracy.
- Improved compliance monitoring accuracy by 16% compared to manual methods.
- Compliance monitoring job was done manually, so the cost was brought down by 18% as the model helped in compliance monitoring.
- A machine learning model made for optimizing card limits based on card usage and expenditure patterns.
- Conducted thorough data analysis and pre-processing to enhance the quality of input data, resulting in more robust models.
- Maintained regular communication with customers to get a better understanding of user behavior patterns.
- Extensively tuned hyperparameters of algorithms to improve accuracy of model.
- Enhanced the user experience which boosted positive reviews by 8%.
- Resulted in significant impact in cost savings of up to 12%.
- A Deep Learning model able to identify facial emotions of input image data. Worked on this project as it was a part of my NIELIT Deep Learning course. Used CNN with 'Relu' as the activation function for each layer and included Average pooling for better accuracy.
- A simple Machine Learning model able to predict if a patient is diagnosed with cancer or not. Used various techniques and algorithms to amplify accuracy of the model.
- A Machine Learning model able to predict if a patient would suffer a heart stroke in future. Used multiple algorithms to compare outputs of each algorithm.