🎓 Senior Computer Science student at The University of Texas at Dallas

🤖 Passionate about Artificial Intelligence, Machine Learning, and Software Engineering

I enjoy building intelligent systems that solve real-world problems. My interests span machine learning, large language models (LLMs), software engineering, and AI-powered automation, with a focus on developing solutions that bridge research and real-world applications.

Currently, I'm a Break Through Tech AI Studio Fellow, where I work on industry-sponsored AI projects, and a former Software Engineering Research Assistant at UT Dallas, where I contributed to research on AI-assisted software engineering and co-authored the CoCoMUT publication.

I'm always exploring new technologies, building projects, and looking for opportunities to collaborate on impactful AI and software engineering solutions.

Developed static analysis pipelines using the SootUp framework to extract software context, generate call graphs, and identify Method Under Test (MUT) relationships for AI-assisted software engineering research.

Built Java and Python workflows to generate high-quality datasets and contributed as a co-author on the CoCoMUT research submission.

Tech Java • Python • Maven • SootUp • Static Analysis

➡️ Research Paper: CoCoMUT (arXiv)

🔒 Repository: Private (research code is not publicly available)

Developed a quantitative analysis platform for exploring financial market data, evaluating investment performance, and generating data-driven insights through interactive visualizations and statistical analysis.

Highlights

• Processed and analyzed historical financial datasets
• Built interactive dashboards to visualize market trends and portfolio performance
• Applied quantitative metrics to support investment analysis and decision-making

Tech Python • Streamlit • FastAPI

➡️ Repository: link

Designed and developed an AI-assisted pet care planning system that generates personalized daily schedules while adapting to real-world constraints such as recurring tasks, scheduling conflicts, pet-specific needs, and owner availability.

Unlike a traditional task planner, PawPal+ uses Retrieval-Augmented Generation (RAG) to actively influence scheduling decisions. Retrieved knowledge adjusts task priorities, flags potential safety concerns for senior pets, and provides context-aware recommendations to improve the quality and reliability of generated schedules.

Highlights

• Built a multi-step agentic planning workflow (input → retrieval → planning → conflict detection → explanation)
• Implemented behavioral RAG that modifies scheduling decisions instead of only generating explanations
• Added recurring task management, conflict detection, and system health monitoring
• Developed automated tests and runtime logging to improve system reliability and transparency

Tech Python • Streamlit • RAG • JSON • pytest • AI Workflow Design • Software Testing

➡️ Repository: link
🎥 Demo: https://www.loom.com/share/c90c57a9680a42c6ad94e947678e73d2

• Python
• Java
• C/C++
• SQL
• JavaScript

• scikit-learn
• TensorFlow
• OpenAI API
• pandas
• NumPy

• React
• Git
• GitHub
• Maven

📧 Email: ajayavardhini@gmail.com

💼 LinkedIn: linkedin.com/in/jaya-vardhini-akurathi

Outside of technology, I enjoy challenging myself through new experiences and creative hobbies. Whether I'm playing badminton or pickleball, baking, or building miniature kits, I enjoy activities that encourage problem-solving, patience, and continuous learning.