Chiranjeevi B S

Logo

Welcome to my portfolio website

View My GitHub Profile

Hi there👋🏼

I'm Chiranjeevi, a passionate software developer with a profound interest in Artificial Intelligence and Deep Learning. Currently, I am pursuing a Master's degree in Computer Science and Engineering at IIT Tirupati. My enthusiasm for technology drives me to constantly learn and innovate, and I find great joy in solving complex problems.


Socials:

You can find me on:

  1. LinkedIn
  2. Medium
  3. Quora
  4. Research Gate
  5. Instagram
  6. Leetcode
  7. Credly
  8. CodeWars.

Areas of Interest


Education


Projects

The Reaper Bot

Description:

Developed an automated Discord bot integrated with Google's Gemini API to enhance user interactions and data analysis. The bot offers: Website Summaries, Web3 Compliance Analysis, Image Analysis, and Interactive Chat. The project utilises gemini-1.5-pro as the text model and gemini=pro-vision as the image analyser model.

Technologies Used: Discord Developer Tool, Python, Flask, Render, Google Generative AI, Prompt Engineering

Link: TheReaperBot


COBOL RuleForge

Description:

Developed a framework for extraction and summarization of Business Rules from legacy COBOL codebases. Generated dataset of COBOL codes and corresponding business rules using few-shot prompting with Google’s gemini-1.0-pro model. Fine-tuned LLMs like Gemma, Llama2, and Mistral on the generated dataset, enhancing the accuracy and quality of the extracted business rules compared to their non-fine-tuned versions

Technologies Used: Python, Jupyter Notebook, Google Generative AI, Mistral AI, Google Gemma, Llama2, Prompt Engineering, Fine-tuning

Link: COBOL RuleForge


Susceptibility of Adversarial Attacks on Medical Image Segmentation Models

Description:

Led an investigative study into the vulnerability of popular medical image segmentation models to white box adversarial attacks like FGSM. Utilized advanced deep learning techniques to assess the robustness of segmentation algorithms against targeted adversarial manipulation. Expanded upon the existing framework by conducting experiments and investigating potential vulnerabilities using an alternative neural network architecture, increasing the attack success rate by almost 6%.

Technologies Used: Python, Jupyter Notebook, PyTorch, PIL, Autoencoders, UNET


EcoTrack

Description:

Developed a Trash Tracker mobile application with Android Studio in alignment with Sustainable Development Goal (SDG) #11 - sustainable cities and communities. Enabled the application to allow users to photograph litter or waste in their communities, geotag the locations, and report incidents along with the specific trash categories for improved local cleanliness and sustainability efforts. Leveraged Google Maps API to integrate real-time tracking and navigation functionalities within the application, enhancing user experience and facilitating efficient reporting of litter or waste incidents

Technologies Used: Android Studio, Java, Firebase, Google Maps API

Link: EcoTrack


Skills


Experience

Intern

SGT Global Technologies India, 03/2023 - 04/2023


Publications

A Comparative Analysis on the Effectiveness of GAN Performance (07/2023)

Abstract:

Generative Adversarial Networks (GANs) have emerged as a potent framework in the discipline of Artificial Intelligence (AI) for generating realistic synthetic data. With the increasing interest and advancements in GANs, there is obligation for a detailed comparative study to comprehend the competencies and vulnerabilities of different GAN variants. This paper sets forth a comprehensive study and comparison of various types of Generative Adversarial Networks (GANs) and their performance in generating high-quality images. GANs have gained popularity in recent years due to their capacity to generate realistic synthetic images. However, the effectiveness of GANs varies depending on the architecture and parameters employed. We have evaluated and compared the performance of different types of GANs, including DCGAN, SRGAN, and CGAN, on benchmark datasets such as CIFAR-10 and MNIST. The evaluation metrics include image quality, standard GAN loss functions and Fréchet inception distance (FID). The results demonstrate that the performance of GANs is highly dependent on the dataset and architecture used, with no single GAN type dominating across all datasets. This comparative study serves as a valuable resource for researchers and practitioners in AI, providing a foundation for selecting the appropriate GAN architecture for specific generative modeling tasks.

Publication Link: A Comparative Analysis on the Effectiveness of GAN Performance

Weather Prediction Analysis using Classifiers and Regressors in Machine Learning (01/2023)

Abstract:

Weather predictions are essential as they protect both property and human lives. Forecasts based on temperature and precipitation are critical for agriculture for merchants in the commodities markets. Humans are also closely related to weather forecast like from daily traveling to daily commuting. Utility firms use temperature projections to predict demand in the upcoming days. Machine learning is a technique that can be used to forecast many different weather patterns, including storms, hurricanes, temperature changes, cyclones, and floods. This study has compared the performance of different deep learning methods for predicting the weather and temperature, including decision tree classifier and SVR, which are rarely used for this purpose. The key variables that affect weather forecasting include air temperature, air pressure, humidity, cloud cover type, solar radiation, precipitation amount and type, and wind speed and direction. Our work focuses on implementing five different machine learning models on a classification dataset and four machine learning models on a regression dataset. Based on our experiments, the regression dataset contributes more towards accurate weather prediction using Decision Tree Regressor, which yields a regression score of 95.70%.

Publication Link: Weather Prediction Analysis using Classifiers and Regressors in Machine Learning


Positions of Responsibility

Class Representative

IIT Tirupati, 08/2023 - Present

Teaching Assistant for the course, CS209M & CS209L : Computer Organisation and Architecture

IIT Tirupati, 01/2024 - 05/2024

Teaching Assistant for the course, ES103M : Introduction to Programming

IIT Tirupati, 08/2023 - 12/2023

Class Representative

Dr. Ambedkar Institute of Technology, 08/2019 - 06/2023


Achievements