Aspiring Software Engineer
Passionate about building innovative software solutions and leveraging machine learning to create impactful applications. Currently pursuing a degree in Computer Science at the University of Texas at Dallas.
End-to-end application development from concept to deployment using modern frameworks and best practices.
Experience in training LLM's with few-shot learning and various calibration technqiues
Building intelligent applications with machine learning capabilities and data-driven insights.
Mentoring developers, leading technical initiatives, and fostering collaborative engineering cultures.
Aesthetic web-based journaling app designed for note-taking. Allows users to journal, track moods, and personalize their experience through theme selection.
Evaluating the davinci-002 model on how accurate it is on P(T) as mentioned by the P(True) calibration experiment from Language Models (Mostly) Know What They Know (Kadavath et al., 2022)
A full-stack web application that leverages GPT-4 and Google Calendar APIs to simplify and automate your scheduling tasks. The app allows users to sign in securely via Google OAuth2, then enter natural language descriptions of meetings or events they want to schedule. Using GPT-4, the app parses these descriptions into structured event details — such as title, duration, preferred time ranges, and attendees.
It then accesses your Google Calendar to check for scheduling conflicts and suggests optimal available time slots within your preferred timeframe. You can select a suggested slot to automatically create the event in your calendar.
• Migrated 7 legacy database tables to modern infrastructure, reducing overall project workload by 50% and
enabling team to focus on high-impact development initiatives
• Modernized Java code affecting 3,000+ JUnit test cases, ensuring compatibility across a large test suite
• Collaborated in an Agile development environment, implementing automated CI/CD pipelines and Git version
control workflows to streamline code deployment and maintain code quality standards
• Evaluated multiple OpenAI models (Davinci-002, GPT-3.5, GPT-4o) using P(True) calibration methodology,
processing 20,000 examples to measure confidence accuracy
• Developed a Python evaluation framework integrating OpenAI API with few-shot learning, temperature
sampling, and k-sampling techniques to optimize automated model assessment, achieving 40% higher
confidence scores
• Implemented statistical analysis and data visualizations using matplotlib and pandas to identify calibration
trends and establish reproducible evaluation protocols
• Developed code for text generation and completion tasks using transformer-based language models, exploring
tokenization, attention mechanisms, and fine-tuning
• Built experimental projects involving Generative Adversarial Networks to explore synthetic text and data
generation techniques
• Implemented and analyzed transformer architectures to understand their application in real-world NLP tasks
I'm always interested in discussing new opportunities, innovative projects, or just connecting with fellow developers. Feel free to reach out through any of these channels.