Ml Engineer Machine Learning

Yuvaraj Reddy Sanagala

AI/ML ENGINEER

OH, USA 551-***-**** **********************@*****.*** SUMMARY

− 3+ years Experienced AI/ML Engineer, Proficient in AI/ML frameworks including TensorFlow, Keras, PyTorch, Scikit- learn, XGBoost, LightGBM, CatBoost, and Hugging Face, with expertise in deep learning architectures such as CNNs, RNNs, Transformers, and GANs for applications.

− Extensive experience in data preprocessing, feature engineering, and feature selection utilizing tools like Pandas, NumPy, SciPy, and Statsmodels, alongside data visualization using Matplotlib, Seaborn, and Plotly to derive actionable insights.

− Skilled in cloud platforms (AWS - Sage Maker, Lambda, EC2; Microsoft Azure - ML Studio, Azure Databricks) and big data tools (Apache Spark, Kafka, Hadoop, Dask, Hive), enabling efficient processing and scalable solutions.

− Hands-on experience in model deployment using Docker, Kubernetes, MLflow, TensorFlow Serving, and Flask/Django, with expertise in version control (Git) and CI/CD workflows. Proficient in hyperparameter tuning through Grid Search, Randomized Search, and Bayesian Optimization.

− knowledge of NLP (Text Processing, Sentiment Analysis, NER with spaCy, NLTK) and Computer Vision (Image Classification, Object Detection, YOLO), with strong capabilities in model evaluation (Precision, Recall, F1 Score, ROC- AUC, Cross-validation, A/B Testing).

SKILLS

Programming & Development Python, R, Java, SQL, Bash scripting, OOP principles, Multi-threading, Parallel computing, Software development (Agile/Scrum methodologies) Machine Learning & Frameworks TensorFlow, Keras, PyTorch, Scikit-learn, XGBoost, Light, CatBoost, Hugging Face, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Transformers, GANs.

Data Science & Analysis Pandas, NumPy, SciPy, Matplotlib, Seaborn, Plotly, Statsmodels, Data Cleaning, Normalization, Feature Engineering, Feature Selection Cloud & Big Data AWS (Sage Maker, Lambda, EC2), Microsoft Azure (ML Studio, Azure Databricks), Apache Spark, Kafka, Hadoop, Dask, Hive Deployment & Operations Docker, Kubernetes, MLflow, TensorFlow Serving, Flask/Django for model APIs, Version Control (Git, GitHub, GitLab, Bitbucket), CI/CD, Hyperparameter Tuning

(Grid Search, Randomized Search), Bayesian Optimization. Specialized Applications Model Evaluation (Precision, Recall, F1 Score, ROC-AUC), Cross-validation, A/B Testing, NLP (Text Processing, Sentiment Analysis, NER, spaCy, NLTK), Computer Vision (Image Classification, Object Detection, Semantic Segmentation, YOLO), Time Series Forecasting, Anomaly Detection. EXPERIENCE

Komodo Health AI/ML ENGINEER, USA May 2023 – Present

− Developed and deployed NLP models for clinical document processing using Python, spaCy, and NLTK, enabling automatic extraction of medical entities such as diseases, medications, and patient histories from unstructured clinical text, achieving 85% accuracy in entity extraction.

− Implemented Hugging Face Transformers (e.g., BERT, RoBERTa) for Named Entity Recognition (NER) and text classification tasks, enhancing clinical document categorization by 80% in terms of precision and recall.

− Preprocessed and structured clinical data using Pandas and NumPy, leading to a 70% improvement in data processing efficiency and preparing data for model training and evaluation.

− Used Scikit-learn and XGBoost for training and fine-tuning text classification models, achieving a 75% improvement in clinical document extraction accuracy through effective feature selection and engineering.

− Leveraged Azure ML Studio and Azure Databricks for scalable training of NLP models, reducing model training time by 80% and accelerating deployment pipelines for clinical document processing.

− Designed and developed a Flask API to expose NLP models for real-time document processing and predictions, enhancing user interaction and enabling clinicians to analyze documents with a 90% reduction in processing time.

− Utilized Azure Kubernetes Service (AKS) for containerizing and orchestrating Python-based NLP pipelines, resulting in 85% improvement in system scalability and fault tolerance during production deployments.

− Automated model deployment with CI/CD pipelines using Azure DevOps, reducing model deployment time by 70% and improving operational efficiency across development cycles.

− Conducted Hyperparameter Tuning with Grid Search and Randomized Search using Scikit-learn, resulting in a 15% improvement in model accuracy for medical term extraction.

− Ensured compliance with HIPAA and healthcare data privacy standards by utilizing Azure Key Vault for encryption and Azure Blob Storage for secure data storage, ensuring 90% adherence to security best practices. HCL Tech AI/ML ENGINEER, India May 2020 – Jul 2022

− Developed and optimized fraud detection models using Java and Apache Spark for distributed data processing, achieving a 75% improvement in transaction processing speed and detection accuracy.

− Implemented machine learning algorithms with Weka and Deeplearning4j, using Gradient Boosting and Neural Networks, improving fraud detection precision by 80%.

− Integrated real-time data streaming using Apache Kafka and Kafka Streams, ensuring 90% real-time processing efficiency for financial transaction data ingestion and anomaly detection.

− Leveraged Java and JDBC to design efficient database connections and queries, optimizing real-time data retrieval and analysis, improving data access times by 70%.

− Applied model evaluation metrics including Precision, Recall, F1 Score, and ROC-AUC, resulting in 85% increase in model performance across different fraud detection cases.

− Designed and implemented Spring Boot-based RESTful APIs to serve fraud detection models in real-time, ensuring 95% uptime and minimizing service latency.

− Utilized Docker for containerization and Kubernetes for orchestration, enabling 80% faster deployment cycles and seamless scaling of fraud detection services in cloud environments.

− Deployed fraud detection models on AWS EC2 instances with AWS Lambda for serverless execution, resulting in a 60% reduction in cloud infrastructure costs while maintaining high scalability.

− Automated model retraining and deployment pipelines using Jenkins and GitLab CI/CD, leading to a 70% reduction in model deployment time.

− Collaborated in Agile/Scrum sprints, contributing to the delivery of iterative improvements and performance optimizations, reducing backlog by 65% through continuous team collaboration and feedback.

− Integrated real-time monitoring and logging frameworks like Log4j and Prometheus, enhancing fraud detection system transparency and uptime by 80%.

− Ensured GDPR-compliant data handling and secure encryption practices, reducing risk exposure and ensuring 90% adherence to regulatory standards for handling financial data. CERTIFICATION

- Predictive Modeling using SAS Enterprise Miner 14.

- Python Basic Certification.

- SQL Advanced.

PROJECT

Predictive Model for Term Deposits

− Utilized a Bank Telemarketing dataset to develop a predictive model for forecasting customer term deposits.

− Employed Python machine learning algorithms and libraries including SKLearn, Imblearn, feature engine, and Pandas for data creation, validation, and modification.

− Generated business insights by identifying key factors influencing term deposit outcomes, Enhanced model accuracy and reliability through comprehensive data processing and feature engineering. EDUCATION

Master of Science in Information Systems: Cleveland State University Cleveland, Ohio, United States. Bachelor of Technology in Mechanical Engineering: Velagapudi Ramakrishna Siddhartha Engineering College, India.

Contact this candidate