Vipul Bhansali
Lead AI Engineer at Sherlocq
India
6+ years of work experience
Sep 20262 mosPresent
Hybrid
Building Sherlocq.ai
Nov 20242 yrsPresent
Bengaluru, Karnataka, India
Building Sherlocq.ai – AI-powered compliance platform.
Jan 20231 yr 11 mosNov 2024
Pune District, Maharashtra, India
1. Developed and deployed ML models using AWS SageMaker, Kubernetes driving a 30% revenue increase and improving call center efficiency by 15%, while implementing customer segmentation techniques to achieve a 60% growth in other product revenue through targeted strategies and CLV analysis.
2. Implemented distributed training and parallel processing using SageMaker and Ray, reducing training time by
40% and iteration speed by 30%, accelerating model deployment and optimizing resource utilization.
3. Leveraged Kubernetes (EKS) to deploy ML pipelines using Kubeflow and Docker containers, enabling customized workflows and reducing operational costs by over 50% while maintaining high scalability and performance.
4. Automated complex ML workflows using SageMaker Pipelines, Step Functions, EventBridge, and Lambda,
reducing manual touchpoints by 90%. This end-to-end automation improved efficiency, minimized errors, and
accelerated deployment cycles, resulting in a more scalable and responsive system.
5. Worked on a POC for an open-source workflow using Ray , MLflow, Airflow. Deployed the solution using Docker and kubernetes, integrating CI/CD pipelines for seamless updates and deployments.
6. Integrated Model and Data Monitoring jobs in the pipeline, enabling proactive detection of performance issues, which resulted in more accurate loan targeting, streamlined telecaller efforts, and minimized business risks.
7. Leveraged AWS SNS for real-time updates to ensure smooth operations by improving stakeholder visibility.
Conducted A/B Testing to set business thresholds, Used shadow deployment to select the best model version,
improving decision accuracy and reducing misclassifications, leading to better business outcome.
Aug 20211 yr 5 mosDec 2022
Pune District, Maharashtra, India
1.Analyzed datasets using Excel, SQL, and Python, integrating Power BI for data visualization.
2. Performed comprehensive (EDA), feature engineering, and feature selection on Mortgage-backed securities datausing Azure Databricks. optimizing model inputs and reducing data processing time by 20%.
3. Built and trained machine learning models to predict prepayment risk for MBS using algorithms like Logistic Regression and XGBoost. Through careful feature selection and hyperparameter tuning improved predictionaccuracy by 15%, allowing BNY Mellon to better manage bond portfolios and reduce exposure to early prepayments, ultimately enhancing profitability and risk mitigation strategies.
4. Developed interactive Power BI dashboards to visualize key metrics and trends related to prepayment risks and model performance, enabling stakeholders to gain real-time insights and make informed decisions for managing mortgage-backed securities portfolios effectively.
5. Implemented advanced data science methodologies, including ensemble learning and feature importance analysis, to enhance model robustness and interpretability. Collaborated with cross-functional teams to integrate these models into business processes, driving strategic initiatives and improving risk mitigation strategies.
Aug 20191 yr 7 mosFeb 2021
Pune
Vipul’s AI Stack
Education
Bachelor of Engineering - BE, Electronics and Telecommunication
Sinhgad college of engineering
20144 yrs2018
Join Vipul and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss