Portfolio

SKILLS & TECHNICAL EXPERTISE

Programming Languages: Python, Java, R, SQL
ML/AI Frameworks & Libraries: TensorFlow, PyTorch, scikit-learn, Keras, NumPy, Pandas, Matplotlib
Data Processing & Tools: Google Colab, Jupyter, Anaconda, MySQL, SQLite, Excel (Pivot Tables, VLOOKUPs)
Core Competencies: Machine Learning, Deep Learning, Computer Vision, Natural Language Processing, Data Visualization, Statistical Analysis
Cloud & Deployment: AWS, Google Cloud, Azure (basic knowledge)

Optimized AI model performance by streamlining data collection, cleaning, and preprocessing, improving data quality by 30%.
Enhanced customer engagement by managing and automating workflows in Odoo CRM, reducing manual intervention by 40%.
Developed and deployed an AI-powered chatbot, improving customer response time by 50% and increasing user interactions.
Automated data extraction using Python web scraping scripts, reducing research time by 60% and enhancing lead generation.

Enhanced model performance by performing data cleaning and optimization, leading to a 20% improvement in efficiency.
Designed and implemented regression models, leveraging statistical analysis and feature engineering to improve predictive accuracy.
Trained and evaluated machine learning models using scikit-learn, optimizing hyperparameters to increase model accuracy and robustness.
Developed insightful data visualizations, enabling data-driven decision-making and improving interpretability of model outputs.

Achieved 98% accuracy in detecting DeepFake videos using CNNs, Vision Transformers (ViT), and BlazeFace.
Improved processing speed by 94% with CUDA-based acceleration for real-time analysis.
Developed an automated Twitter bot using Selenium to detect, analyze, and respond to DeepFake verification requests.
Trained on Facebook’s DFDC dataset with over 120,000 videos for robust detection.
Designed for scalability and cloud deployment on AWS and Google Cloud for efficient large-scale processing.

Increased image classification accuracy to 95% using Support Vector Machines (SVM), Deep Learning, and Neural Networks.
Optimized image processing speed by 40%, reducing classification time from 5 seconds to 3 seconds per image.
Automated dataset collection, expanding training data by 30%, using Bing Image Downloader for large-scale image acquisition.
Enhanced feature extraction efficiency by 50% with advanced preprocessing techniques in NumPy, Matplotlib, and scikit-learn.
Reduced misclassification errors by 35% through iterative model training and fine-tuning of hyperparameters.
Scalable deployment-ready model, designed for real-time image recognition on cloud-based platforms.

The Digital Looking Glass: Predicting DeepFake Evolution through Social Media Bot Analysis, ZKG International vol. IX, 2024
- Issue I.DOI: https://bit.ly/DigitalLookingGlass

Osmania University

Bachelor Of Engineering - Computer Science (Specialized in AI & ML) Hyderabad, India July 2024 CGPA: 7.90
Relevant Courses: AI, Advanced ML, Data Mining, DBMS, Big Data, Deep Learning

IEEE Member (2021-2024): Assisted in event organization and inventory management.