Praneeth Sunkavalli Email & Phone Number

Malvern, PA, US

• Project: Enhancing Cybersecurity Readiness in Non-Profit OrganizationsAwarded as the First recipient of the Presidential Strategic Initiative on Public Impact Research (PPIRA)
• Scraped data from 80,000 non-profit organizations to extract financial information such as revenue, assets, and organizational types. Developed and implemented metrics to evaluate and measure the social impact of these.
• Created a solution to filter non-profit organizations to 400 by leveraging cluster analysis and assessing criteria like revenue, assets, and tax form types, enabling the precise identification of relevant organizations.
• Analysed survey data (descriptive questions and interviews), applying NLP techniques with the T5 LLM model to summarize the responses and quantifying information to categorize the NPOs into safe, at-risk, and.

• Project: Influence of Societal Readiness on Energy Marketplace
• Developed a BeautifulSoup-based tool to gather the past 4 decades’ financial information of multiple energy sector companies, including energy production and consumption, and analyzed public sentiment towards nuclear.
• Constructed an OLS model to assess the impact of societal acceptance of sustainable energy on the Market Capital and Returns of energy companies achieving an R-squared value of 0.7 while accounting for energy.
• Conducted a thorough analysis to determine the relationships between staff expenses and energy production and consumption, as well as between return volatility and these energy metrics, resulting in a linear.

• Project Title: Inflation Expectations Inferred from Economic News
• Extracted 40 years’ worth of public information on the Federal Rate and Consumer Price Index, subsequently managing 10,000+ data points on a MySQL server database.
• Implemented NLP and pre-trained embedding models such as Universal Sentence Encoder to transform the textual data into a numeric vector, followed by classification models to establish an AUC of 0.73 to predict the.
• Evaluated diverse embedding techniques (tiktoken, sentencepiece) and deep learning algorithms including Transformers and advanced NLP models (BERT, GPT ), leading to a significant increase in model accuracy from 60% to.

Pennsylvania, United States

• Project: High-Frequency Trading Forecast and Adversarial Attacks
• Programmed a multivariate time series deep learning architecture, integrating CNN for feature extraction, coupled with an LSTM model, achieving an accuracy of around 78% and an F1-score of 0.7 on high-frequency trading.
• Conducted gradient-based and optimization-based adversarial attack algorithms with varying epsilon values to evaluate model resilience, revealing a significant average accuracy reduction of 42%.
• Managed and sustained Docker containers to ensure version control and streamlined deployment for 3 teams.
• Implemented KPI dashboards and reports to monitor model performance, attack parameters, and security metrics, enabling initiative-taking adjustments and enhancements.

• Provided mentorship and guidance to a 7-person undergraduate team throughout the development of predictive models for the American Red Cross, meeting all specifications for a year-long analytics project.
• Secured and integrated supplementary data for model optimization, including critical details such as livelihood information, population density, weather details, and many more, which raised the accuracy by 7%.
• Experienced in implementing various machine learning and deep learning models across diverse domains, focusing on optimizing model efficiency. Proficient in conducting data research to identify and integrate.
• Conducted bi-weekly progress meetings with students and coordinated updates with the partnered companies.
• Partnered with multinational companies such as Microsoft, Accenture, and Lockheed to support prototype development and problem-solving and leveraged industry insights and resources to support the project.

Hyderabad, Telangana, India

• Developed 10+ sustainable pipelines facilitating the seamless deployment of scalable machine learning models from research to production on the Google Cloud Platform.
• Discovered crucial trends and patterns using statistical analytics and modeling, which enhanced strategic decisions.
• Collaborated closely with 4 teams from various fields to analyze their workflows and suggest deep learning models that substantially improved accuracy and efficiency in data processing, thereby enhancing operational.

Hyderabad, Telangana, India

• Engineered machine learning models with an accuracy of over 70%, and diagnosed to generate crucial insights that drive business decisions and organizational goals.
• Created around 30+ detailed reports, 20+ engaging dashboards using Tableau, and visually appealing presentations to communicate data-driven information effectively. This helped convey complex information clearly and.
• Cross-functional collaboration to achieve project goals, demonstrating strong communication and teamwork skills.

Master'S Degree, Data Analytics, 3.96/4.0

Bachelor'S Degree, Computer Science, 8.7/10

• For the final year project, we made a subscription model for the stock price prediction, using NBeats, CNN, LSTM, and FBProphets..