Prompt Manipulation with Large Language Models (LLMs)

Document Summarization

Summarize lengthy documents, highlighting key concepts and essential information to enhance comprehension and accessibility.

JavaDoc Generation

Automate the generation of Javadoc, streamlining the documentation process by producing accurate and consistent comments based on code analysis.

Integration Test Generation

Automate the generation of Java integration tests, enhancing test coverage and improving development efficiency.

Iterated Prompt Manipulation interacting with LLMs

JUnit Test Generation

Execute AI-generated JUnit tests in Java, providing error feedback to the model for automatic test repair and iterative refinement.

Project Reverse Engineering

Analyze software projects to offer a detailed breakdown of their structure and functions, enabling accurate insights for project analysis and improvement.

Develop Machine Learning Models

Text Classification
Document categorization

Applied SVM for document categorization by training on labeled data, enabling the model to accurately predict categories for new documents.

Image Classification
Food recognition

The application identifies food items in uploaded images and provides the user with various recipes that can be made using those ingredients.

Regression
Empoyee attrition prediction

One particularly interesting project involved building a machine learning model designed to predict employee attrition, helping identify which employees might leave the company.

Creating Serverless Endpoint on Cloud Provider

LLAMA Model Service Provider

Deploying a serverless endpoint on a GPU-powered cloud provider to efficiently host and scale our LLAMA model.

Text Analysis
Quantitative Text Analysis and Linguistic Statistics

By leveraging statistical data such as word and sentence counts, readability indices like Flesch and Gulpease, and quality parameters like lexical density and sentence complexity, we can evaluate the clarity and coherence of textual data.

Local ML model on the mobile device
Handwritten Recognition

Inference of a local model on the mobile device for document identification and text recognition.