Vishal Tyagi

Detail-oriented Machine Learning enthusiast with a strong analytical mindset and a passion for developing innovative AI-driven solutions. Possesses a solid foundation in machine learning, deep learning, and data analysis, backed by hands-on experience through academic projects and internships. Adept at problem-solving and continuously exploring new technologies to enhance model performance. Demonstrates effective collaboration within teams while also excelling in independent research and implementation.

Research Paper Publication

• Title: Survey Paper on Multilingual Live Call Translation using Deep Learning Journal: Journal of Image Processing & Pattern Recognition Progress (stm journals)
• Authors: Vishal Tyagi, Divyansh Agrawal, Vibhu Sehgal, Nishant Shokeen, Nishi Gupta
• Abstract: This paper surveys advanced language translation technologies, exploring the mechanisms of multilingual, real-time translation in voice calls.
• Publication Details: Vol. 11, Issue 2, Pages 13-21, 2024

Phishing Detection Using Machine Learning & Deep Learning (Python)

• Developed a phishing detection system using a stacked ensemble of Random Forest, XGBoost, and 1D CNN, achieving 96.38% accuracy.
• Applied Select-k-Best for feature selection to enhance efficiency and reduce noise.
• Performed 10-fold cross-validation for robust evaluation and reduced false positives.
• Optimized hyperparameters using Optuna, benchmarking superior performance against traditional models.
• Utilized the UCI Phishing Websites Dataset for training and evaluation.

Face Recognition-Based Attendance System (Python)

• Built a real-time attendance system using Python, OpenCV, and the 'face recognition' library.
• Enhanced face recognition under varying lighting conditions using CLAHE (Contrast Limited Adaptive Histogram Equalization).
• Implemented face detection, encoding, and time-stamped attendance marking for accurate record-keeping.
• Integrated Google Cloud Datastore for storing face encodings and attendance data.
• Developed a Tkinter-based GUI for adding new faces and managing attendance records.
• Ensured accuracy and efficiency with a cooldown feature to prevent duplicate entries.

Live Stock Price Prediction Using Machine Learning

• Developed a real-time stock price prediction system utilizing machine learning techniques.
• Implemented models such as Random Forest, XGBoost, and LSTM to analyze historical stock data and predict future prices.
• Integrated sentiment analysis of financial news and social media data to enhance prediction accuracy.
• Optimized model performance using feature engineering and hyperparameter tuning.
• Deployed the system with a user-friendly interface for real-time stock trend visualization.

Volunteer Manager & Social Media Head, Ek Bharat Shreshtha Bharat, Northcap University, Gurugram, Haryana

• Managed volunteer coordination and social media strategies to promote cultural exchange initiatives.
• Organized and executed campaigns to enhance engagement and participation in university programs.

Volunteer, Sarv Sangi, Vikaspuri, New Delhi

• Participated in food donation drives and initiatives to support underprivileged children through education.

Volunteer, Various Clubs & Societies, Northcap University, Gurugram, Haryana

• Engaged in multiple social activities, including plantation drives, cleanliness awareness programs, and donation drives.
• Contributed to community service initiatives organized by university clubs.