nikhil = {
    "role"       : "B.Tech CSE (AI/ML) — University of Mumbai (2023–2027)",
    "focus"      : ["Machine Learning", "Computer Vision", "Cybersecurity", "NLP"],
    "currently"  : "Building production ML systems | Phishing detection + YOLOv8",
    "ambassadors": ["C-DAC Mumbai", "Google Student Ambassador", "GfG Campus Mantri"],
    "internships": ["Code B Solutions", "Edunet Foundation", "Oasis Infobyte"],
    "trained"    : "50+ students in AI/ML technologies",
    "mission"    : "Fighting cybercrime with ML & promoting India's tech 🇮🇳",
    "status"     : "Open to AI/ML Internships | Placement-ready for 2027"
}

Languages & Core

Machine Learning & Data Science

Computer Vision & Deep Learning

Deployment & DevOps

Development Tools

Production-ready cybersecurity ML system | 89.63% accuracy

• Analyzed 11,430 URLs with comprehensive EDA (histograms, correlation heatmaps, pair plots)
• Built Random Forest classifier achieving 89.63% accuracy on 57 URL-extractable features
• Implemented SHAP explainability — identified google_index & special_char_ratio as critical
• Deployed via Flask REST API with modern dark-themed UI & confidence scoring
• Dockerized for production deployment across any platform
• Real impact: Detects phishing URLs in <100ms with visual confidence bars

YOLOv8-powered detection | 29 objects in single frame

• Detected 29 objects simultaneously with 92% confidence on complex scenes
• Dual modes: Image upload analysis + Live webcam streaming through browser
• Built with YOLOv8 (COCO 80-class detection) + Flask backend
• Real-time bounding box visualization with class labels & confidence scores
• Clean modern UI with detection statistics dashboard

NLP-based text classifier | TF-IDF vectorization

• Built spam classifier using Natural Language Processing techniques
• Applied TF-IDF vectorization for intelligent feature extraction from text
• Trained on real-world spam dataset with high precision/recall balance
• Identifies spam patterns: urgency words, suspicious links, poor grammar

Smart agriculture | Flask web app

• ML model recommending optimal crops based on soil NPK, pH, temperature, humidity, rainfall
• Helps farmers maximize yield through data-driven crop selection
• Flask deployment with clean HTML interface for field data input
• Real-world impact: Sustainable farming through precision agriculture

Regression model | Used car valuation

• Predicts fair market value based on brand, year, mileage, fuel type, transmission
• Comprehensive EDA revealing depreciation patterns across segments
• Insights: Diesel cars depreciate slower, automatic commands 30-40% premium

Advertising ROI analysis | Budget optimization

• Forecasts product sales based on TV, radio, newspaper advertising spend
• Key finding: TV advertising shows strongest ROI, newspaper minimal impact
• Provides actionable budget allocation recommendations for marketers

Key Metrics:

• 🔥 40+ total contributions since Aug 2024
• 📈 Current streak: 18 days
• 🎯 6 production-ready projects deployed
• 💻 Primary language: Python (62.2%)
• 📦 11 repositories created

• 🏅 AI/ML Learning Track — Gen AI Academy 2.0
• 🏅 Naukri Campus Young Turks 2025 — Selected participant
• 🏅 Build Real World AI Apps — Google (Gemini & Imagen)
• 🏅 Data Analytics Job Simulation — Deloitte Australia

"India doesn't just need to consume global technology — it needs to build its own."

My contribution to Indian tech ecosystem:

🛡️ Cybersecurity: ML-powered phishing detection (89.63% accuracy) protecting users from online fraud

🎯 Computer Vision: Real-time object detection systems (29 objects/frame @ 92% confidence)

🇮🇳 Indigenous Tech Advocacy: Promoting DHRUV64 microprocessor & Bhashini AI to 50+ students

🌾 Sustainable Agriculture: Smart crop recommendation system for data-driven farming

👥 Community Impact: Trained 50+ students in practical AI/ML applications

Machine Learning

• Supervised Learning: Random Forest, Decision Trees, Linear/Logistic Regression, SVM
• Model Evaluation: Cross-validation, hyperparameter tuning, confusion matrices
• Explainability: SHAP values, feature importance analysis
• Feature Engineering: Selection, extraction, normalization, encoding

Computer Vision

• Object Detection: YOLOv8 (80 COCO classes)
• Real-time Processing: Webcam streaming, frame-by-frame analysis
• Image Operations: OpenCV preprocessing, augmentation, transformations

Natural Language Processing

• Text Classification: Spam detection, sentiment analysis
• Feature Extraction: TF-IDF vectorization, n-grams
• Preprocessing: Tokenization, stemming, stop word removal

Data Science

• EDA: Pandas, Matplotlib, Seaborn visualization
• Statistical Analysis: Correlation studies, distribution analysis
• Data Cleaning: Missing values, outliers, normalization

MLOps & Deployment

• Web Frameworks: Flask REST APIs with HTML/CSS/JS frontends
• Containerization: Docker for reproducible deployments
• Version Control: Git/GitHub for collaborative development
• Production Pipelines: Model serialization (pickle), scalers, preprocessors