Denton, Texas, USA

When a large language model confidently returns a wrong answer to a high-stakes clinical or financial query, who bears the cost — and how do we prevent it? This question has defined my research trajectory. Working at the intersection of machine learning and real-world data infrastructure — from building financial data pipelines at scale to studying the failure modes of generative AI — I have come to focus on a single, urgent problem: how do we build AI systems that know what they don’t know? I am fascinated to develop scalable frameworks for uncertainty quantification and trustworthiness in AI systems operating on high-volume, real-world data.

• Reliability and Trustworthiness of AI Systems
• Machine Learning and Statistical Learning Theory
• Scalable Data Systems and Large-Scale Data Analytics
• Uncertainty Quantification in Predictive Models

University of North Texas — Master of Science in Advanced Data Analytics
May 2024 • GPA 3.636 / 4.0

Gokaraju Rangaraju Institute of Engineering and Technology — B.Tech in Electronics and Communication Engineering
May 2022 • GPA 3.36 / 4.0

Graduate Research Affiliate — Reliability of Large Language Models
Computational Healthcare & BioTechnology Lab, Dr. Mohammed Aledhari, University of North Texas • Jan 2025 – Present

• Investigating failure modes and uncertainty-aware decision-making processes in large language models (LLMs) to address the critical problem of confident yet inaccurate AI outputs.
• Designing experimental protocols to evaluate model calibration and reliability across diverse query domains, contributing to the development of trustworthy AI evaluation frameworks.
• Analyzing patterns of model hallucination and overconfidence, identifying systematic failure categories that inform uncertainty quantification strategies for real-world deployment.

Capstone Research — Energy Consumption Forecasting
University of North Texas • Fall 2023 – Spring 2024

• Built predictive machine learning pipelines to forecast U.S. energy consumption using large-scale time series datasets, implementing preprocessing, feature engineering, and anomaly detection workflows.
• Evaluated multiple model architectures using RMSE and MSE metrics, revealing how temporal dependencies and distributional shift degrade prediction accuracy in long-horizon forecasting.
• Produced research-style technical reports and visualizations supporting sustainable energy planning applications.

Transportation Safety and Spatiotemporal Analytics Project
National Student Data Corps • Jan 2025

• Conducted end-to-end geospatial and temporal analysis of large-scale NYC traffic collision data, including schema validation, missing-value diagnostics, and bias-aware preprocessing.
• Applied time-series decomposition, seasonality extraction, anomaly detection, and geospatial hotspot analysis to identify temporal risk patterns and spatial crash clusters.
• Synthesized findings into a professional research poster presented to the U.S. Department of Transportation Federal Highway Administration.

Independent Research — Structural and Demographic Patterns in Custodial Arrests
University of North Texas • May 2024

• Performed large-scale exploratory data analysis on Dallas Police Department arrest records using statistical aggregation, categorical encoding, and spatial pattern detection.
• Uncovered non-intuitive patterns in high-density residential environments and weapon involvement prevalence, generating evidence-based insights for municipal resource allocation policy.

• B. V. Kumar, A. Bharat, G. V. Eathamokkala, Y. S. S. Harsha, A. U. Sree, "Analysis of an IoT based Water Quality Monitoring System," I-SMAC 2022. DOI: 10.1109/I-SMAC55078.2022.9987360.
• Al-Edhari, A., G. V. Eathamokkala, & Rahouti, M. (2026). Response drift across frontier large language models. Manuscript under review at Nature Machine Intelligence.

• Spatiotemporal Analysis of New York City Traffic Crash Data — Research Poster, National Student Data Corps; presented to USDOT FHWA, Jan 2025.

Data Engineer — Vsion Technologies, Austin TX • Sep 2024 – Present

• Design scalable data pipelines integrating Kafka streaming with PostgreSQL analytical storage layers for large-scale structured financial datasets, identifying computational bottlenecks that motivate research in scalable data systems.
• Develop optimized relational data models and layered analytical views enabling downstream statistical analysis and machine learning workflows, with emphasis on reproducibility and experimental consistency.
• Implement feature engineering pipelines and apply query optimization and modular schema design to improve computational efficiency for large-scale ML deployment.

Data Analyst — Zetatek Technologies Pvt Ltd, Hyderabad, India • Jan 2022 – Dec 2022

• Analyzed operational and financial datasets using SQL Server and SSIS to identify statistical trends and optimize resource allocation strategies across business units.
• Developed automated analytical dashboards and integrated SQL-based data pipelines for structured reporting and visual analytics communication.

TriSQL Framework — Text-to-SQL Research Implementation

Independently implemented a three-stage Text-to-SQL framework inspired by the TriSQL architecture (Nature Scientific Reports, 2026), converting plain English questions into executable SQL queries using open-source tools running entirely on local hardware.

• Designed and built a semantic schema selector using sentence-transformers (all-MiniLM-L6-v2) to filter relevant database tables via cosine similarity, reducing prompt noise and improving generation quality.
• Developed a two-step structured SQL generator that first identifies required SQL clauses (JOIN, GROUP BY, WHERE) before generating the complete query — improving syntactic correctness over single-prompt approaches.
• Implemented a complexity-aware refinement stage that classifies generated SQL as Easy, Medium, or Hard and applies tiered error correction including execution feedback loops for hard queries.
• Evaluated on the Spider benchmark dataset (Yale University) — achieving 70% Execution Accuracy and 100% Executability Rate using SQLCoder via Ollama with no GPU or API costs.
• Deployed a FastAPI web interface enabling non-technical users to query any SQLite database in plain English and view results directly in a browser.

GitHub: https://github.com/Gowthamch9/trisql-framework

IoT-Based Water Quality Monitoring System (E-Aqua)

• Engineered an IoT-enabled prototype integrating pH, turbidity, TDS, temperature, and flow sensors with an Arduino-based controller for real-time environmental data acquisition.
• Implemented wireless data transmission via ESP8266 Wi-Fi module, enabling continuous remote monitoring through a cloud-connected dashboard and mobile interface.
• Conducted comparative analysis of monitoring technologies, demonstrating a low-cost, scalable system for municipal, aquaculture, and agricultural applications. Published in IEEE.

• Programming and Data Science: Python (NumPy, Pandas, scikit-learn, PyTorch, TensorFlow), SQL, PySpark, R
• Machine Learning and Statistics: Regression, Classification, Clustering, Neural Networks, Time Series Forecasting, PCA, Random Forests, Support Vector Machines, Cross-Validation, Model Evaluation, Uncertainty Quantification
• Big Data and Cloud Systems: Apache Spark, Kafka, Google Cloud Platform (BigQuery, Dataflow, Pub/Sub, Dataproc, Cloud SQL, Vertex AI), Snowflake
• Databases: PostgreSQL, MySQL, Microsoft SQL Server
• Visualization and Tools: Matplotlib, Seaborn, Power BI, Tableau, Git, Jupyter Notebook, LaTeX

• Coding Tutor: Taught Python and Scratch to middle and high school students through project‑based learning (~25 students).
• Volunteer Data Science Mentor: National Student Data Corps — guided teams on preprocessing, analysis, and poster development.

Certifications:

• Google Cloud Data Engineering and ML Specialization
• Google Advanced Data Analytics Professional Certificate
• Microsoft Power BI for Data Analysts
• Advanced SQL for Data Engineering.

Honors:

• Selected Participant, National Student Data Corps
• Graduate GPA 3.636 / 4.0, University of North Texas.

• English: Professional proficiency
• Telugu: Native
• Hindi: Conversational proficiency
• Tamil: Conversational Proficiency