Y. Eswar

M.Tech Data Science candidate · AI/ML, Data Science & Analytics

Hyderabad, Telangana, India

@eswaryadav23452413

LinkedIn GitHub

0 followers

🎓 M.Tech in Data Science at GITAM School of Technology · Graduating 2027

About

M.Tech Data Science candidate and B.Tech CSE (AI) graduate building projects across machine learning, software, and analytics. Experienced with LLM fine-tuning, anomaly detection APIs, computer vision, desktop app development, and Power BI reporting.

Education

2027

M.Tech

GITAM School of Technology

Data Science · 2027

Data Science
2025

B.Tech

Parul Institute of Engineering and Technology

Computer Science & Engineering (AI) · 2025

Computer Science & Engineering (AI)

Skills

<3 months

C SQL .NET LLMs NumPy Flask Swift Python Pandas OpenCV PyTorch SwiftUI TensorFlow Data Science scikit-learn Computer Vision Machine Learning API Integration Understanding Generative AI (ChatGPT, Copilot, Gemini)

Tools / apps / platforms

<3 months

Git MySQL GitHub Vercel MongoDB Google Colab Microsoft Power BI Hugging Face Transformers

Projects

Power BI Analytics - Rapido & Superstore Sales

Power BI, data analysis

Built Power BI analyses of ride and sales data to surface trends for business questions.
Miracle - Open-Source Local File Converter

Swift/SwiftUI, C#/.NET, GitHub

Built a desktop app for Mac and Windows that converts files entirely on the user’s machine with nothing uploaded; the Windows build passed all 303 feature checks.
Image Colorization - Collaborative Development

Python, PyTorch, TensorFlow, Git

Co-developed a U-Net model to colorize grayscale CIFAR-10 images using peer code review and iterative feedback.
Few-Shot Breast Histopathology Classification

Python, PyTorch, ResNet18, Google Colab

Reproduced and extended a Prototypical Networks approach (ResNet18 backbone) for image classification with limited labelled data, testing pretrained encoders, feature fusion and classifiers under patient-separated evaluation.
Llama 3.2 Fine-Tuning for Finance Question Answering

Python, PyTorch, Hugging Face Transformers, LoRA/QLoRA

Adapted Llama 3.2 to finance-domain queries using LoRA/QLoRA; compared adapted outputs with baseline responses to assess domain accuracy and response quality, and documented methodology and findings.
EduSentinel - Educational Anomaly Detection

Python, PyTorch, Flask, REST API

Converted an open-ended anomaly detection problem into an end-to-end pipeline and deployed a real-time inference REST API, carrying the project from analysis through implementation. Achieved 90.43% accuracy and 99.40% ROC-AUC through structured evaluation across multiple training runs; documented methodology and validated results.

📚 Publications

Abnormal Learning Behaviour Detection Using Deep Autoencoders
Real-Time Face Detection ↗

Explore the Broxer community

💬 Q&A Ask about roles, salaries and interviews — and help others. 👥 Circles Join discussions on skills, companies and industries. 📰 Feed Follow updates and insights from the community.