Sr. Data Scientist
i2c
Total des années d'expérience :7 years, 5 Mois
● Developed and deployed machine learning models to detect and prevent fraudulent activities for over 1M cardholders in real-time, resulting in a reduction of fraud loss by 40%.
● Conducted thorough data analysis on millions of transactions to identify patterns and trends, leading to the development of a new fraud detection algorithm that achieved an f1 score of 89 percent.
● Collaborated with cross-functional teams including IT and Risk Management to implement a real-time alert system that automatically flags suspicious transactions, reducing response time by 50%.
● Dealt with data imbalance and data integrity problems.
● Data preprocessing, labeling processes, cleaning and extracting meaningful information from the data to be used as features.
● Handled categorical features and feature scaling.
● Performed out of time validations and evaluating model performances.
● Identifying data drift and concept drift in models in production to estimate model decay beforehand.
● Advising the customer on the use of AI, capturing the requirement and translating the requirement into suitable mathematical models.
● Analyze the data and client's requirements and suggest a data driven solution of
business challenges.
● Developed and executed machine learning algorithms to improve image recognition accuracy by 15%, resulting in a decrease of false positives by 25% and a savings of $250k in manual review costs.
● Find the avenues/points of maximum opportunity and evaluate them in terms of dollars or dollar return of that decision.
• Suggest actionable steps/ insights that can realize that dollar value and make
sure those insights should be in alignment with overall organizational goals, brand image goals and goals of specific department that are going to implement these actionable insights.
● Implemented deep learning models for customer demand forecasting, resulting in a 20% improvement in accuracy compared to traditional time series models and driving revenue growth of $1M.
● Identified company’s sales performance across dimensions like regions, product categories, and customer segments.
● Identified interesting customer segments considering purchased products and affinity for discounts.
● Identified customer segmentation problem.
● Work with the other data scientists on understanding their ML models.
● Built API endpoint for interacting with ML models.
● Worked on pickle and mleap for model result conversion.
“aghhh love usman! 10/10 would recommend” (Source: Fall 2012 anonymous course evaluation)
Le lien a été supprimé pour non-respect des conditions d'utilisation. Veuillez contacter l’équipe d'assistance pour plus d'informations.