Data Science Machine Learning

RAMESWARI MISHRA

Email: ad5wtf@r.postjobfree.com Cell: 573-***-****

LinkedIn: https://www.linkedin.com/in/rameswarimishra/ SUMMARY

Data Science MS student with industry experience in data science and software development. Seeking full-time opportunities in Data Analysis or Software engineering and AI/Machine Learning modeling. Open to relocation EDUCATION

Master of Science, Data Science and Analytics, University of Missouri, Columbia Expected: July 2024 (3.95/4) Master of Technology, Computer Science and Engineering, NIST, Biju Patnaik University of Technology, Odisha, INDIA, 2018 (8.8/10) Bachelor of Science, Computer Science and Engineering, RIT, Biju Patnaik University of Technology, Odisha, INDIA, 2006 TECHNICAL SKILLS

• Languages: Python, R, SQL, Java, C++, MATLAB, JavaScript

• Big Data/Cloud: MySQL, AWS (S3, EC2), PostgreSQL, Kubernetes, Docker, Azure

• Data Science, AI, ML Algorithms: Large Language Models (LLMs), Generative AI, Machine Learning, Deep Learning, Natural Language Processing (NLP), Explainable-AI, algorithm development, statistical analysis, supervised and unsupervised learning, Classification, Clustering, Hypothesis Testing, Manova, Developing Machine Learning Pipelines

• Data Science Tools: Keras, TensorFlow, PyTorch, Scikit-learn, SciPy, NumPy, Pandas, Seaborn, Matplotlib, Plotly, ggplot2, PowerBI, Tableau

• Other Skills: Software Testing, CI/CD pipelines, Image Processing, ETL pipeline, Relational DBMS, Data cleaning, and Visualization

• Other Tools: Excel, Jupyter, Git, RStudio, JIRA, Figma, Google Cloud Platform, Hadoop PROFESSIONAL EXPERIENCE

Data Science Intern, Johnson County Mental Health Center, Misson, Kansas, Jan 2024 – May 2024

• ML project deployment collaboration:

o Collaborating with MLOps and other teams to execute ETL Pipeline and deploy ML projects addressing Behavioral Health Crises, supporting 50 highly vulnerable patients each month.

o Tracking Git versions, logs, reports, and metadata to provide actionable insights on interventions, code fixes, model details, and results for teams.

o Leveraging insights from EHR data analysis with ML algorithms and Tableau to develop effective intervention measures proactively. Data Science Intern, Blessing Health in Quincy, IL, May 2023 – Aug 2023

• Patient Panel Adjustment:

o Developed predictive model pipelines with precise hyperparameter tuning and grid search, visualizations, and reports to adjust patient panels for providers using Electronic Health Records data, within deadlines. o Conducted data curation, cleaning, and feature engineering for multi-class classification models. o Achieved 0.72 test set accuracy with XGBoost, a 15% improvement over state-of-the-art literature.

• Model Deployment Collaboration:

o Collaborating with the MLOps team to deploy the model into production.

• Visualizations using Tableau dashboards and scorecards: o Took the initiative to publish a work hour tracker tableau dashboard, using timesheets, enhancing productivity by 20% via informed decision-making on work allocation and resource optimization for the Data Science Team. o Launched patient appointments and panel distribution tracker tableau dashboards and scorecards using PostgreSQL views, aiding management decision-making through monthly reviews. Automation Engineer, Inditek Pioneer Solutions, Mar 2021- May 2022

• Automated data quality evaluation for fertilizer data using Python that improved productivity by 15%, and reduced processing time by 50%.

• Implemented CI/CD pipeline and automated email reporting system, enhancing customer engagement and increasing business value by 15% with Java Script and Jasmin Framework.

Software Engineer, Thrymr Software, Nov 2018 - Mar 2021

• Developed ML models and data visualizations (Python and R libraries) for strategic decision-making improved productivity by 9%.

• Collaborated in 12 agile projects with national and international clients, in complete Software Development life cycles, ETL and Automation. RESEARCH EXPERIENCE

Graduate Research Assistant, Center to Stream HealthCare in Place, University of Missouri, Columbia, Nov 2022 – Present

• Project: Reducing False Alarms in Fall Detection using Deep Learning Models: o Developing deep learning models (ConvNet with LSTM, vision-transformer) to detect real-time falls in apartments using depth-video data. o Exploring multi-modal LLMs for improved detection accuracy and adding context to fall alerts, implemented containerization with Docker & utilizing S3 Bucket for data storage. Validated robust human detection in-depth videos with 99.9% accuracy, utilizing a YOLOv5 model on a sample test comprising 100 random annotated frames, evenly split between 50 falls and 50 non-falls. Graduate Researcher, National Institute of Science and Technology, Odisha, India, July 2017- April 18 Thesis Title – “Image Inpainting using Deep Learning Techniques.”

• Developed Denoising Stacked Auto Encoders using Keras and TensorFlow in AWS. Data Science Course Projects, University of Missouri, Columbia

• Lung Cancer Prediction:

o Built a decision tree-based lung cancer prediction model with 0.94 accuracy using Kaggle Survey Lung Cancer data. o Conducted feature importance analysis to identify key predictors.

• Effect of COVID-19 Vaccination Analysis and Visualization: o Conducted comparative trend analysis of vaccination effects in Alabama, Missouri, and NY. o Demonstrated reduced COVID-19 deaths and infection rates with increased vaccination rates.

• Heart Failure Survival Prediction:

o Developed logistic regression model for heart failure survival prediction (accuracy: 0.82) using UCI Heart Failure dataset. o Conducted feature importance analysis using Manova.

• Supply chain optimization:

o Implemented a system to predict and prevent back orders, achieving an F1 score of 0.95 for non-back ordered items. o Achieved 0.81 recall for predicting back-ordered items.

• Multi-year Chicago Crime Database Design, ETL, and Analytics o Engineered ETL pipelines in PostgreSQL for Multi-Year Chicago Crime Database, ensuring data integrity, cleanliness, and normalization. o Conducted trend analysis on crime and arrest counts utilizing PostgreSQL queries and Python visualizations, revealing a consistent decline in total crime count and a widening disparity between crime occurrences and arrest rates.

Contact this candidate