As a leader in Data Science and Cyber Specializations at UCNJ, I am committed to fostering academic excellence and advancing education in the field of Data Science. My role integrates instructional expertise, program coordination, and student engagement to ensure comprehensive support for learners.Key responsibilities include curriculum development, leveraging technical expertise, aligning programs with industry standards, fostering outreach and collaboration, and implementing strategic support systems to guide students toward success. These efforts are designed to prepare students for meaningful careers in both academic and industrial settings immediately upon graduation, equipping them with the knowledge and skills needed to excel.

○ Newly appointed role focusing on the use of high-performance computing (HPC) to manage and optimize software solutions for bioinformatics applications.○ Responsibilities include configuring and maintaining software environments tailored for bioinformatics research, ensuring high levels of computational efficiency and data processing capabilities.○ Collaborate with academic and research staff to develop and implement new software tools that enhance the capabilities of bioinformaticians in data analysis and research.

○ Developed and launched a full-stack Flask-based educational application tailored for stock market enthusiasts, featuring interactive tools for stock analysis, educational resources, and secure user authentication. The application also includes a financial news article summarizer using a BART transformer, and a forecasting tool integrating an LSTM-transformer with sentiment analysis based on 4 API calls, aimed at predicting stock trends. Explore the application (without ML models) at vasemili.pythonanywhere.com.○ Thoroughly tested and implemented several machine learning algorithms for stock market forecasting, with a particular focus on LSTM-transformers, achieving an R2 score of 0.9338 and MAE of 0.0108 for a 5-day forecast on average across 10 different companies’ datasets. Enhanced forecasting accuracy through Federated Learning and sentiment analysis using NewsAPI and the FineBERT model.○ Pioneered the development of an automatic tradebot by integrating advanced trading techniques with transformerbased algorithms, substantially enhancing trading strategies and decision-making processes in real-time financial markets.

○ Developed interactive data visualizations for cost distributions using APIs, enhancing the usability of cloud cluster monitoring for the data engineering team. Integrated these visualizations into a full-stack website built with Flask, MongoDB, JavaScript, HTML, and CSS, improving user accessibility and insights into data usage. ○ Analyzed and interpolated data from cluster usage and spending, identifying inefficiencies and estimating an annual cost savings of approximately $10,000 by optimizing cluster operations. Developed an automated alert system to notify the data engineering team of long-running clusters, leading to more efficient resource management. ○ Contributed to the development of scalable data pipelines that ingested large volumes of data from Kafka Topics, processed it using Azure Databricks, and stored results in Google Cloud Storage (GCS). This streamlined data processing workflow supported real-time decision-making and analytics. ○ Deployed and managed Docker images of the web application using Azure pipelines and Kubernetes, ensuring a robust and scalable deployment process. This enhanced the application’s performance and reliability, aligning with best practices in software development and deployment.

○ Organized and cleaned federal university data, enhancing data accuracy and usability.○ Successfully migrated the university’s databases from Denodo to MS SQL Server, transferring functions, operations, views, aggregations and tables, documenting each step using Python. This documentation was later given to the new organization in charge of Carnegie Classifications, the American Council on Education for better visualization and understanding of classifications.○ Conducted PCA analysis on Carnegie Classifications, identifying critical factors for R1 and R2 institutions. We also presented the categorical changes of Universities and their categories in flows using Sankey Diagrams

○ Directed a project analyzing COVID-19 vaccination rates using machine learning algorithms (LR, Lasso, Ridge, ElasticNet, DTR, SVR, KNR, RFR and GBR). Utilized machine learning analytics to identify characteristics of communities with lower vaccination rates, providing valuable insights: Medical history and disease susceptibility, Wealth and education, Families with children, and Co-living populations explain an impressive 67% of variation in unvaccinated counts among ZIP codes of California.