Parth Vadera

Built scalable machine learning pipelines for personalized recommendations and job matching using modern frameworks such as PyTorch, TensorFlow, and MLflow, improving user engagement and interaction rates. Deployed real-time inference systems using cloud-native tools and stream processing platforms, reducing latency and ensuring consistent performance for millions of daily users. Designed and main

Led AI risk detection architecture including fraud blocking, trust scoring, and suspicious activity mining to ensure platform security at scale. Developed and deployed a deep learning fraud detection model for financial transactions, preventing $12M+ in fraud, and replacing a third-party solution. Engineered the next-generation recommendation engine using advanced transformer architectures and mul

Led the development of global demand forecasting models at the product level using advanced machine learning techniques (PyTorch, XGBoost, SQL), enabling data-driven planning across regions and improving forecasting accuracy. Designed and deployed loyalty propensity models using deep learning with embedding layers in Keras and TensorFlow, resulting in a 12-point year-over-year increase in redempti

Developed advanced attribution models using statistical and deep learning techniques to evaluate the effectiveness of large-scale advertising campaigns on digital engagement and web traffic trends. Built automated tools for exploratory data analysis and anomaly detection on third-party datasets using distributed computing frameworks. Implemented model training, deployment, and monitoring workflows

Collaborated with healthcare data teams to develop generative classification models using statistical and machine learning techniques, uncovering patterns between risk factors and mental health outcomes. Led a research initiative funded by a national science organization to analyze and simulate referral pathways in healthcare systems using probabilistic modeling and network analysis. Explored earl

Used Pandas, NumPy, scikit-learn to develop predictive models and perform customer segmentation analysis, driving data-informed marketing decisions and increasing overall campaign effectiveness. Developed time series forecasting models using Holt-Winters smoothing to predict daily website traffic, supporting better planning of advertising budgets and server resource allocation. Contributed to the

Focused on Operation Research, gaining expertise in quantitative methods and analytical techniques to solve complex problems. Applied statistical and mathematical models to optimize processes and decision-making.

Studied Information and Communication Technology, building a strong foundation in computer science and communication systems. Gained practical skills in various aspects of IT and its applications.