Portfolio Layout

About Me

I am Jayanth Devaraj Gowda I am a graduate Student pursued a degree in Computer Science Engineering and Data Science.I am passionate about leveraging technology to solve real-world problems and have a strong foundation in Python,Machine learning and data analysis.My goal is to continuously learn and grow as a data scientist, contributing to projects that create meaningful impact.

One of my key areas of interest is machine learning, and I've worked extensively with Python, using libraries such as TensorFlow, Opencv, Scikit-learn, and Pandas. My focus has been on building models that can make data-driven predictions, automate tasks, and derive insights from complex datasets. In addition to machine learning, I’m also passionate about data analysis, which involves exploring and visualizing data to uncover hidden patterns and trends. I have developed these skills through various academic projects and personal initiatives, which have given me hands-on experience in solving real-world problems.

I am deeply interested in how AI can be applied to make better decisions and create more efficient systems. Ultimately, I aspire to be part of a team that pushes the boundaries of technology and makes a positive impact on society."

Resume

EDUCATION

ATME College Of Engineering 2021 - 2025, Mysuru, Karnataka

Bachelor's Degree in Computer Science and Data Science, Visvesvarya Technological University

CGPA : 8.86/10

Gopalaswamy Shishuvihara Independent PU College2019 - 2021, Mysuru, Karnataka

PCMB,Karnataka Sate Board

88%

Rotary High School2016 - 2019, Mysuru, Karnataka

10th Standard,Karnataka Sate Board

87.16%

EXPERIENCE

Generative AI Developer InternFebrauary 2025 - May 2025

Synthlinx Private Limited Bengalru, Karnataka

Integrated open-source LLMs (Llama, Mistral, Falcon) and APIs (OpenAI, Gemini, Groq, Claude, Qwen) for real-world applications.
Built RAG (Retrieval-Augmented Generation) pipelines using FAISS / ChromaDB for knowledge-based AI responses.
Developed AI-powered chatbots, code assistants, or creative content tools using Flask & Fast API for deployment.
Applied fine-tuning and prompt engineering to optimize generative model outputs for specific tasks.

Machine Learning InternNovember 2023 - Decemeber/2023

Anurva Advanced Techserv Pvt Ltd Mysuru Karnataka

Analyzed large datasets to identify trends and patterns.
Developed machine learning models to predict the outcomes.
Collaborated with the team to implement data-driven solutions.
Developed and implemented machine learning models for classification and regression tasks.
Cleaned and preprocessed large datasets to optimize model performance.
Evaluated model performance using metrics like accuracy, F1-Score, MAE, MSE and RMSE.

My Skills

TECHNICAL SKILLS

Machine Learning
Generative AI
Data Vizualization
Deep Learning
Agetic AI
Agetic Framework
LLM
RAG Models
Pandas& numpy
VectorDB,ChromaDB,MongoDB
Neural Network
Data Base Management
Cloud Computing
Streamlit
TensorFlow & scikit-learn
Opencv & Keras
Data Analytics
Tableau and Power BI

CODING

80%

PYTHON

90%

HTML&CSS

70%

SQL

75%

JAVA

85%

C

35%

ANGULARJS

60%

ReactJS

DESIGN

Web Design

Sketching

Photoshop

3D Modeling

LANGUAGES

English

ಕನ್ನಡ (Kannada)

हिंदी (Hindi)

Recent Works

FACE EMOTION DETECTION

The project involves creating a system that can automatically recognize and classify human emotions from face expressions captured in images or videos using Machine Learning Models.

RAITHA MITHRA WEBSITE

The project involves creating a system like Crop Recommendation, Yield and Price Prediction, and Disease Classifier integrated into a single interactive webpage using ML and Streamlit.

SPEECH EMOTION DETECTION

The project involves creating a system that can automatically recognize and classify human emotions from audio-speech in videos & display the emotion in a dashboard using ML Models.

CRICKET WIN % PREDICTOR

The project involves creating a system that can predict the win percentage of cricket matches using machine learning models, analyzing historical match data, player statistics, and other relevant factors.

VIDTUBE

The project involves creating a YouTube clone using HTML, CSS, and JavaScript. It offers a sleek and responsive user interface, mimicking the look and feel of the original platform.

1] FACE EMOTION DETECTION

Project Overview:

Dataset: FER2013 Labeled Grayscale Images
Libraries: OpenCV, TensorFlow, Keras, scikit-learn, Seaborn, pandas, numpy
Algorithms: Haar Cascade Algorithm, Convolutional Neural Network
Input: Normal Image or Video Capturing
Output:

Happy
Sad
Angry
Surprised
Fearful
Neutral

Specifications:

The CNN picks out the important patterns and find simple patterns like edges or corners and ignores the less important details. It focus on the key features of the face(eye movements, mouth shape, eyebrow position, nose and cheeks). Pattern-finding process happens in multiple layers. Each layer looks for more complex patterns, like shapes and textures, by combining the simpler patterns found in earlier layers After several layers CNN decides what emotion is being shown in the face (like happy, sad, or angry) based on the patterns it has learned

IMAGES

2] RAITHA MITHRA WEBSITE

Project Overview:

Dataset: Soil, Weather, Historical Crop yield and Market Price Dataset
Libraries: OpenCV, TensorFlow, Keras, streamlit, scikit-learn, Seaborn, pandas, numpy
Algorithms:

Specifications:

This project provides a comprehensive decision-support system for farmers by integrating crop recommendations, yield and price predictions, and disease diagnostics into a single platform. By utilizing machine learning and real-time data, the system helps farmers make informed decisions about what crops to plant, anticipate yields, and maximize profits by understanding price trends. Additionally, early disease detection allows for timely interventions, reducing crop losses. The interactive nature of the Streamlit webpage makes it easy for users to input data and receive actionable insights, enhancing agricultural productivity and sustainability

IMAGES

3] SPEECH EMOTION DETECTION

Project Overview:

Dataset: RAVDESS Dataset
Libraries: OpenCV, TensorFlow, Keras, scikit-learn, Dash, Librosa, pydub, Seaborn, pandas, numpy
Algorithms: MFCCs, CNN, RNNs
Input: Video Files, Live Camera with Microphone
Output: Emotions visualized with bar charts and waveforms using a Dashboard

Specifications:

The System is designed to analyze and identify emotions from speech in video files, system extracts the audio and extracted audio is then processed to extract features like MFCCs using Librosa, which are fed into a deep learning model built with TensorFlow or Keras to classify emotions. The results are displayed on an interactive dashboard which showcases the original video alongside visual representations of emotion detection, including bar charts for emotion distribution and waveforms for emotion intensity over time, The System enables real-time emotion analysis from videos, which can enhance customer service by understanding emotional tones in interactions, assist therapists in assessing patients' emotional states for better diagnosis and treatment, and improve entertainment experiences by tailoring content based on viewer emotions.

IMAGES

4] CRICKET WIN% PREDICTOR

Project Overview

Dataset: Previous Cricket Matches Dataset
Input: Team's Current Score, Wickets and Number of Overs
Output: Win% of the Team

Specifications

The system is designed to forecast the probability of a team winning a cricket match by analyzing historical data, player statistics, and other relevant factors. It utilizes machine learning models to process data such as past match results, player performance metrics, pitch conditions, and team compositions. By applying algorithms like Logistic Regression, Random Forest, and Gradient Boosting, the system predicts win percentages and provides insights into match outcomes. This predictive capability aids teams, analysts, and enthusiasts in making strategic decisions and understanding match dynamics. The system offers a data-driven approach to enhance decision-making and strategic planning in cricket.

IMAGES

5] VIDTUBE

Project Overview and Specifications

The project involves creating a web application that mimics the core functionalities of the popular video-sharing platform. Users can browse, upload, and watch videos, create playlists, and interact through comments and likes. The application features a user-friendly interface with video playback capabilities, a search function for discovering content, and a recommendation system to suggest videos based on user preferences. It also includes user account management for personal profiles and watch histories. The system is built using technologies like HTML, CSS, JavaScript, The project aims to provide a similar user experience to YouTube, with the added benefit of customization and scalability based on specific requirements.