█████╗ ██████╗ ███╗ ██╗ █████╗ ██╗ ██╗ ██████╗ ██╗ ██╗██████╗ ████████╗ █████╗
██╔══██╗██╔══██╗████╗ ██║██╔══██╗██║ ██║ ██╔════╝ ██║ ██║██╔══██╗╚══██╔══╝██╔══██╗
███████║██████╔╝██╔██╗ ██║███████║██║ ██║ ██║ ███╗██║ ██║██████╔╝ ██║ ███████║
██╔══██║██╔══██╗██║╚██╗██║██╔══██║╚██╗ ██╔╝ ██║ ██║██║ ██║██╔═══╝ ██║ ██╔══██║
██║ ██║██║ ██║██║ ╚████║██║ ██║ ╚████╔╝ ╚██████╔╝╚██████╔╝██║ ██║ ██║ ██║
╚═╝ ╚═╝╚═╝ ╚═╝╚═╝ ╚═══╝╚═╝ ╚═╝ ╚═══╝ ╚═════╝ ╚═════╝ ╚═╝ ╚═╝ ╚═╝ ╚═╝
arnav = {
"role" : "AI/ML Developer & CS Undergraduate",
"institute" : "Vishwakarma Institute of Technology, Pune",
"focus" : ["Deep Learning", "Transformers", "NLP", "Fraud Detection"],
"published" : "Springer LNNS — ICAIN 2025",
"languages" : ["English", "Marathi", "Hindi", "Japanese (conversational)"],
"beyond_code": ["Multiple Olympiad Medalist (STEM)",
"National Gold Medalist — Skating 🥇",
"Black Belt — Karate 🥋"],
}|
Music Genre Classification · GTZAN Dataset CNN–Transformer hybrid leveraging temporal attention on spectrogram data. Achieved ~95% classification accuracy using spectrogram augmentation and attention-based modeling.
|
Fintech · Uncertainty-Aware Pipeline Uncertainty-aware fraud detection using TabNet, ensemble models, and feature engineering. Optimized for CPU-efficient, scalable deployment.
|
|
Hackathon Project · AI + LLM Rule-based clinical pathway engine integrated with Groq-hosted LLaMA 3.1. Safety-first architecture separates evidence-based rules from LLM explanations.
|
Hackathon Project · Dual-Model Pipeline NLP-based SMS scam analysis (TF-IDF + LR) + transaction risk scoring (Random Forest). ~85–90% accuracy with explainable risk assessments.
|
Adaptive TabNet–β-VAE Fusion Ensemble for Uncertainty-Aware Fraud Detection
Springer Lecture Notes in Networks and Systems (LNNS) · ICAIN 2025
Languages Python · C++
ML / DL PyTorch · Scikit-learn · TabNet · Transformers · CNNs · LLMs · NLP
Web Flask · React
Tools Git · Colab · LaTeX · Linux
CS Core Data Structures & Algorithms · OS · Compilers
"Precision in research. Discipline in execution. Curiosity in everything."