Data Scientist Machine Learning

RAJASHEKAR KORUTLA

+1-857-***-**** *******.*@************.*** Maine, USA Open to relocate linkedin.com github.com SUMMARY

Experienced Data Scientist with 3+ years of experience in developing and deploying predictive models for a variety of industries. Skilled in data analysis, machine learning, and statistical modeling to drive insights from complex datasets. EDUCATION

Northeastern University September 2022 - June 2024 Master's, Applied Machine Intelligence GPA: 3.8

Jawaharlal Nehru Technological University

Bachelor's, Mechanical Engineering GPA: 3.5

PROFESSIONAL EXPERIENCE

THE INSTITUTE OF EXPERIENTIAL AI

NLP Engineer Gen AI Engineer. September 2023 - Present

• Generative AI for Healthcare: Developed a Generative AI model using llama to summarize electronic health records, with an aim to significantly reduce response times and saving medical professionals 4 hours daily.

• Predictive Modeling for Patient Care: Engineered a predictive model with deep learning techniques in Python, achieving an AUC of 0.87 for post- surgery risk assessment. Designed and implemented data transformation processes to enhance patient care management.

• Quality Control in Manufacturing: Led a project that combined machine learning algorithms to predict 3D printed part surface roughness with over 95% accuracy. Environment: Python, R, TensorFlow, PyTorch, Scikit-learn, Hive, MapReduce, Hadoop, AWS EMR, AWS S3, SQL, Docker, Jupyter Notebook, Git

THORNTON TOMASETTI

Data Scientist Machine Learning Engineer July 2023 - December 2023

• Developed an advanced energy modeling tool using EnergyPlus and SMOTE for data balance significantly reducing modeling time while maintaining high accuracy. Utilized Pandas for data manipulation, NumPy for numerical operations, and SciPy for scientific computations.

• Utilized XGBoost algorithms to enhance energy consumption predictions with an R2 of 92% and deployed the model on AWS for scalability. Created user-friendly interface with Flask and Bootstrap for local testing, transitioning to an AWS-hosted cloud solution for improved efficiency.

• Performed A/B tests, ensuring the deployment of the most effective model for operational use.

• Leveraged AWS S3 for storage solutions, SageMaker for model training and deployment, BigQuery for handling large datasets, and Docker for containerization, ensuring a seamless and scalable workflow. Environment: AWS S3, SageMaker, Python(Pandas, NumPy, SciPy, XGBoost), Big Query, Docker, Git, Flask, Bootstrap TATA CONSULTANCY SERVICES

Data Scientist Computer Vision Engineer July 2021 - August 2022

• Defect Detection: Developed and implemented advanced computer vision pipelines using TensorFlow and PyTorch, achieving 95% accuracy in defect detection for manufacturing processes. Employed techniques such as convolutional neural networks (CNNs) and transfer learning to enhance model precision.

• Safety Compliance: Engineered real-time object detection systems for workplace safety monitoring, utilizing YOLOv3 and OpenCV. Developed custom data augmentation strategies to improve model robustness and accuracy under diverse environmental conditions.

• Medical Imaging: Improved diagnostic accuracy with image segmentation using U-Net and ResNet50.

• Deployed high-performance vision models using Docker and Kubernetes to ensure robust and reliable production environments. Conducted extensive A/B testing and continuous integration/continuous deployment (CI/CD) to optimize model performance and deployment efficiency.

• Utilized AWS EC2 and S3 for scalable computing and storage solutions, SageMaker for model training and deployment, and Airflow for orchestrating complex data workflows. Developed interactive front-end interfaces with Flask and React for seamless user interaction and real-time monitoring. Environment: TensorFlow, PyTorch, OpenCV, YOLOv3, U-Net, VGG16, ResNet50, AWS EC2, S3, SageMaker, Docker, Kubernetes, Flask, React, Airflow.

AI TECHNOLOGIES

Junior Data Scientist January 2019 - May 2021

• Data Workflow Optimization: Streamlined data workflows using Python and Hadoop, increasing efficiency by 20% and reducing errors by 40% through the implementation of automated data pipelines and ETL processes.

• Content Personalization: Improved user engagement by 10% and customer satisfaction by 15% by developing traditional machine learning models such as decision trees and logistic regression for content recommendation.

• NLP for Content Discoverability: Enhanced content discoverability using fundamental NLP techniques such as tokenization, stemming, and lemmatization with TensorFlow and NLTK.

• Conducted basic A/B testing to validate the effectiveness of machine learning models. Environment: Python, TensorFlow, Hadoop, OpenCV, NLTK, A/B testing. PROJECTS

• Generative AI for Healthcare: Developed a model to summarize electronic health records using LLaMA 3, BERT, and VLLMs from Hugging Face libraries, enhancing clinical response times with metrics such as ROUGE-1, ROUGE-2, BERTScore, BERT precision, and BERT recall. Access link

• Predictive Modeling for Patient Care: DL model for post-surgery risk assessment with an AUC of 0.92. Access link

• Quality Control in Manufacturing: ML model to predict 3D printed part’s surface roughness with 95% accuracy.

• Advanced Energy Modeling Tool: Reduced energy modeling time leveraging ML for Energy Plus. Access link

• Computer Vision for Safety Compliance: Engineered real-time object detection systems using YOLOv3 and PyTorch.

• Reinforcement Learning Projects: various projects using the RL environments of the gym library. Access link

• Text Classification with NLP: Enhanced content discoverability using transformers and LLMs like BERT, GPT-3, and RoBERTa, alongside embeddings such as Word2Vec, FastText, and GloVe.

• Optimizing Base Locations of LifeFlight of Maine: Analyzed transactional data and identified the closest base locations for quick patient reach using various clustering algorithms. Access link

• Recommender System for LL Bean: Developed a recommender system to suggest similar products when items go out of stock by analyzing transactional and clickstream data and employing encoding techniques and distance calculations from product attributes content personalization. Access link SKILLS

Programming & DevOps: Python, SQL, R, Jenkins, Kubernetes, Terraform, CUDA, GPU Programming, MLflow, DVC, Golang, CI/CD

Databases & Cloud: MySQL, Oracle, Postgres, Snowflake, S3, EMR, EC2, Lambda, Redshift, Sagemaker, Athena, Spark, Hadoop, Kafka, GCP, Azure

ML & DL: Linear, Tree-based, Ensemble methods, TensorFlow, DNN, CNN, RNN, LSTM, Transformers, BERT, GPT-3.5, T5, EfficientNet, MobileNet

Development, Reporting & Version Control: Agile, JIRA, Trello, Tableau, Power BI, Git, Plotly, R Shiny, Flask, React, Apache Airflow, Feature Engineering

RESEARCH AND PUBLICATIONS

• Korutla, R., & Gooje, V. (2024). Enhanced building energy consumption prediction using a dual-model approach and ensemble techniques. Manuscript under review. The Roux Institute, Northeastern University.

• Korutla, R., Correa, M., Zheng, C., & Post, N. L. (2024). Characterizing and predicting as-built surface roughness in LPBF using machine learning. Manuscript in preparation. The Roux Institute, Northeastern University.

• Korutla, R., Hicks, A., Mazhude, F., Kelting, T., Rabb, J. B., Jin, Q., Kramer, R., Sawyer, D., Winslow, R. L., & Amal, S. (2024). Deep learning model for predicting adverse events after cardiothoracic surgery in the ICU using STS data and time series intraoperative data. Manuscript in preparation. The Roux Institute, Northeastern University; Maine Medical Center; Spectrum Healthcare Partners.

• Wendelken, S., Antony, A., Korutla, R., Pachipala, B., Shanahan, J., & Saba, W. (2024). Roux-lette at “Discharge Me!”: Reducing EHR Chart Burden with a Simple, Scalable, Clinician-Driven AI Approach. In Proceedings of the 29th Conference on Medical Image Understanding and Analysis. CERTIFICATIONS

- TensorFlow Professional Developer Certification, DeepLearning.AI

- Deep Learning Specialization, Coursera (offered by DeepLearning.AI)

- Machine Learning Certification, Stanford University, via Coursera

- Citi Certification for Conducting Healthcare Research

Contact this candidate