Pallavi Yadav

I’m an AI/ML engineer with hands-on experience building end-to-end machine learning pipelines, from data processing and EDA to predictive modeling and deployment-ready workflows. I focus on performance, scalability, and practical model monitoring so solutions hold up in real-world settings.

In my internships, I processed 1M+ transactional records using SQL-based ETL pipelines, improving data retrieval efficiency by 30%. I also built interactive Power BI dashboards with 15+ KPIs using DAX, reducing decision-making time by 25% through clearer, faster insights.

My deep learning and applied ML work includes an Urban Flood Risk Prediction system using a CNN-LSTM approach on satellite data, plus an end-to-end early warning pipeline with Google Earth Engine and geospatial processing. I also developed ModelGuard, a pre-deployment failure detection system using PyTorch/TensorFlow across 1000+ model runs, reaching an F1-score of 0.98 for detecting overfitting, data leakage, and instability.

I enjoy spanning domains like computer vision, NLP, and time series/geo-spatial analysis, using tools like MLflow, Streamlit, and Linux to make models easier to diagnose and ship. I’m also Azure AI Fundamentals certified, and I bring an active, team-oriented mindset from workshop coordination and editorial/content roles in AI and renewable energy communities.