Tanvica Samudrala

Fine-tuned RoBERTa and BART transformers on 100K+ nonprofit filings for mission statement classification, achieving F1 of 89% on high-value funder identification Built semantic search system using transformer embeddings to match research topics across 100K+ donor profiles, improving retrieval precision by 40% over keyword baselines. Engineered automated visual analytics pipeline processing 2.5M+ grant records annually, reducing research turnaround time by 90%.

Designed and deployed a production AI research platform on Oracle Cloud (FastAPI, PostgreSQL) supporting 5 experimental conditions and 300+ participants in an active IRB-approved human-AI interaction study Built a coverage-aware LLM routing engine using Groq-hosted models, dynamically selecting among 3 specialized agents with 87% correct routing accuracy and role-specific follow-up probing across all experimental conditions Reduced candidate evaluation time from 45 to 12 minutes per session, saving 110+ researcher hours per semester across active studies.

Promoted brands such as Engage and NewMe through strategic social media campaigns, increasing brand awareness and engagement among the student community.

Standardized 4 years of meteorological sensor data across 75+ districts into a 100K-record ML-ready dataset with 92% accuracy, used for government disaster reporting. Built Tableau dashboards with real-time flood risk and rainfall deviation indicators, cutting manual government reporting from days to hours Applied statistical trend analysis to identify high-risk geographic zones, directly informing disaster preparedness and resource allocation decisions

• On AWS Cloud Foundation & AWS Data Analytics Courses.

Developed and benchmarked CNN architectures with optimized preprocessing pipelines and hyperparameter tuning, improving model reproducibility across experimental configurations. Evaluated deep learning model performance across multiple training setups, contributing to architecture refinement for published experimental analysis.

Graduate coursework and projects focused on machine learning, large language models, knowledge graphs, statistics, and applied AI. Developed advanced understanding of data science essentials, machine learning, and system design. Key coursework: Applied Machine Learning, Introduction to Statistics, Advanced Database Technology, Social Media Mining, Data Visualization, Knowledge Graphs and Large Language Models, Database Design, Information Visualization.

Undergraduate coursework and projects specialized in Data Science along with strong foundations on machine learning, visualizations, neural networks, data mining, operating systems and architecture, and big data analytics.