Srikar Lakkimsetti

About Srikar Lakkimsetti

AI Engineer with 4+ years of experience building production-grade Machine Learning and…

Experience

1. Machine Learning Engineer

   Allstate

   Aug 2024 — Present · Northbrook, IL, US

   Designed machine learning pipelines using Python, Scikit-learn, and XGBoost to predict claim approval probability and fraud likelihood across multiple insurance products, supporting claim assessments per month.Developed TensorFlow models with Keras layers for claim severity estimation and premium adjustment, improving prediction accuracy and optimizing loss ratio evaluations for underwriting teams.Built PyTorch models integrated with TorchServe for near real-time fraud detection and anomaly monitoring, increasing event detection throughput across high-volume claims by 2.3×.Implemented NLP pipelines to extract insights from claim documents, customer feedback, and adjuster notes using SpaCy and NLTK, enabling structured data availability from unstructured text records.Engineered feature extraction workflows with Pandas and NumPy to consolidate policy, claim, and customer data, building a risk evaluation matrix that cut manual audit time from 2 hours to 45 minutes.Applied LLMs such as BERT and GPT-3.5 through fine-tuned transformer models for policy summarization, claim classification, and entity tagging, reducing manual text review workloads across departments.Deployed end-to-end ML workflows on Azure Machine Learning and Azure Data Factory, ensuring automated retraining, version tracking, and secure data integration for production-grade deployments.Collaborated with data engineering and DevOps teams to implement model validation and drift monitoring pipelines, tracking precision, recall, and F1-score metrics to maintain consistency across retraining cycles.

Education

• University of Illinois Springfield

  Master's degree, Computer Science