Aditya K

Data Scientist @Coinbase

US

Developed and deployed a fraud detection model using XGBoost, SHAP, and Apache Spark, reducing fraudulent transactions by 32% while optimizing real-time risk assessment across millions of cryptocurrency transactions per second. Built and maintained a real-time anomaly detection pipeline using Kafka, Flink, and PostgreSQL, improving system monitoring, reducing downtime by 45%, and enhancing security against unauthorized transactions in blockchain networks. Trained, optimized, and scaled large-scale deep learning models with PyTorch and TensorFlow, increasing model accuracy by 28% for financial risk predictions by integrating feature engineering and hyperparameter tuning methods. Designed and deployed an NLP-based financial insights tool leveraging BERT, spaCy, and Named Entity Recognition (NER), automating news sentiment analysis and improving trade recommendations by 22% for high-frequency traders and investors. Orchestrated the migration of ML models to Kubernetes-based deployments, improving inference latency by 38%, increasing throughput, and enabling seamless model updates without disrupting trading activities. Created a personalized recommendation engine using Collaborative Filtering, Graph Neural Networks, and Neo4j, leading to a 19% boost in user engagement by dynamically adjusting trading insights based on behavioral data.

PRASAD V POTLURI SIDDHARTHA INSTITUTE OF TECHNOLOGY

Bachelor of Technology - BTech

Illinois Institute of Technology

Master's degree