Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I am a data-driven AI/ML engineer with a focus on end-to-end machine learning, NLP, and Generative AI solutions. I have worked on insurance analytics, risk modeling, and workflow automation, delivering measurable improvements in decision accuracy and operational efficiency. I thrive in translating complex business problems into deployable AI systems and robust data pipelines. I design production-grade LLM workflows and scalable ML pipelines, leveraging cloud-native architectures with Python, PyTorch, AWS, Snowflake, and LangChain. I integrate model explainability into production solutions and build dashboards to empower enterprise teams with actionable insights.…I am a data-driven AI/ML engineer with a focus on end-to-end machine learning, NLP, and Generative AI solutions. I have worked on insurance analytics, risk modeling, and workflow automation, delivering measurable improvements in decision accuracy and operational efficiency. I thrive in translating complex business problems into deployable AI systems and robust data pipelines. I design production-grade LLM workflows and scalable ML pipelines, leveraging cloud-native architectures with Python, PyTorch, AWS, Snowflake, and LangChain. I integrate model explainability into production solutions and build dashboards to empower enterprise teams with actionable insights.

Tushar Vantaram

Data Scientist, AI Engineer, Full Stack Developer, +3





I am a data-driven AI/ML engineer with a focus on end-to-end machine learning, NLP, and Generative AI solutions. I have worked on insurance analytics, risk modeling, and workflow automation, delivering measurable improvements in decision accuracy and operational efficiency. I thrive in translating complex business problems into deployable AI systems and robust data pipelines. I design production-grade LLM workflows and scalable ML pipelines, leveraging cloud-native architectures with Python, PyTorch, AWS, Snowflake, and LangChain. I integrate model explainability into production solutions and build dashboards to empower enterprise teams with actionable insights.…I am a data-driven AI/ML engineer with a focus on end-to-end machine learning, NLP, and Generative AI solutions. I have worked on insurance analytics, risk modeling, and workflow automation, delivering measurable improvements in decision accuracy and operational efficiency. I thrive in translating complex business problems into deployable AI systems and robust data pipelines. I design production-grade LLM workflows and scalable ML pipelines, leveraging cloud-native architectures with Python, PyTorch, AWS, Snowflake, and LangChain. I integrate model explainability into production solutions and build dashboards to empower enterprise teams with actionable insights.

Available to hire

I am a data-driven AI/ML engineer with a focus on end-to-end machine learning, NLP, and Generative AI solutions. I have worked on insurance analytics, risk modeling, and workflow automation, delivering measurable improvements in decision accuracy and operational efficiency. I thrive in translating complex business problems into deployable AI systems and robust data pipelines.

I design production-grade LLM workflows and scalable ML pipelines, leveraging cloud-native architectures with Python, PyTorch, AWS, Snowflake, and LangChain. I integrate model explainability into production solutions and build dashboards to empower enterprise teams with actionable insights.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

AI/ML Engineer at Liberty Mutual Insurance

January 1, 2015 - November 5, 2025

Engineered advanced ML pipelines by preparing large-scale structured and unstructured insurance data, applying NLP techniques to accident reports and customer descriptions, and leveraging AWS S3 and SageMaker for scalable model training that improved model readiness for production deployment. Designed and optimized fraud risk scoring models using XGBoost and deep learning approaches, integrating domain-specific feature engineering and anomaly detection techniques, which enhanced the adaptability of fraud detection beyond rule-based systems. Collaborated with cross-functional teams including claim investigators and business analysts to align fraud detection insights with compliance goals, deploying containerized APIs with Docker and Kubernetes into the claim management system for real-time scoring. Analyzed claim processing outcomes to identify inefficiencies and implemented automated monitoring dashboards in Power BI that provided transparency into fraud detection precision, directly i

Machine Learning Engineer at Hexaware Techno, India

December 1, 2022 - December 1, 2022

Engineered scalable ML pipelines by processing scanned policy forms, claim reports, and medical records with OCR and NLP, applying data cleaning, entity recognition, and validation workflows to improve information accuracy in insurance document automation. Designed feature engineering strategies combining internal customer histories with external risk indicators, leveraging predictive analytics and deep learning techniques to deliver comprehensive customer risk assessments for underwriting decisions. Deployed containerized APIs with Docker and Kubernetes, integrating ML models into insurance platforms for real-time document processing and risk scoring, while ensuring reliability through continuous monitoring and automated retraining cycles. Collaborated with data scientists, business analysts, and claim investigators to align model outcomes with regulatory compliance, building dashboards in Tableau and Power BI that tracked fraud detection, risk scores, and automation coverage percenta

AI/ML Engineer at Liberty Mutual Insurance

January 1, 2025 - Present

Built enterprise-grade LLM workflows using LangChain, GPT-4, and Pinecone to automate claims summarization and document intelligence, reducing adjuster manual review time by 45%. Designed ML models for premium risk scoring and loss-ratio forecasting using XGBoost and TensorFlow, improving underwriting decision accuracy and reducing false-risk predictions by 21%. Engineered feature pipelines via AWS Glue, Airflow, and Snowflake to support real-time scoring, achieving 97% data freshness across actuarial and claims datasets. Implemented MLOps pipelines using Docker, Kubernetes, and GitHub Actions for automated retraining, drift alerts, and staged deployments, reducing deployment cycles by 40%. Integrated SHAP-based explanations into risk models for compliance, enabling transparent decision trails and accelerating audit acceptance across regulatory teams.

Machine Learning Engineer at Hexaware Technologies

June 1, 2020 - December 1, 2022

Developed ML models for transaction classification, customer segmentation, and fraud-pattern detection using TensorFlow, scikit-learn, and PySpark, improving analytic accuracy across financial workflows. Designed ETL pipelines integrating structured and semi-structured data using PySpark, Hive, and SQL, increasing pipeline throughput and reducing manual preprocessing effort. Built NLP solutions for text extraction and entity tagging using spaCy and BERT to automate document processing, reducing human review cycles for client operations. Automated model deployment and monitoring workflows using AWS, Jenkins, and Airflow, improving reliability and reducing pipeline failures. Partnered with data engineering teams to enhance feature store design and data quality rules, improving model consistency and reducing debugging cycles by 20%. Created Power BI dashboards to visualize risk metrics, customer patterns, and model outputs, enabling faster decision-making for client analytics teams.