I am a graduate student at Cornell University pursuing my Master of Professional Studies (MPS) in Applied Statistics.

Before coming to Cornell, I spent over two years as a Software Engineer at UBS Bank, where I built cloud-native search infrastructure, deployed large-scale microservices on Kubernetes (AKS), and engineered real-time data ingestion pipelines handling millions of records.

My current work bridges the gap between rigorous classical statistics and modern machine learning engineering, with a strong focus on causal inference, Bayesian data science, and LLM applications.

• Causal ML Capstone: Architecting a Double Machine Learning (DML) framework using the PSID dataset to estimate the Average Treatment Effect of job training. I am currently evaluating whether LLM embeddings (MiniLM, DistilBERT) of unstructured household histories can serve as valid proxies for latent confounders.
• Deep Learning Benchmarking: Quantifying spatial inductive bias and sample efficiency by benchmarking CNNs (ResNet) against Vision Transformers (ViT) on progressive data splits.
• Hardware & IoT: Designing and assembling a battery-powered e-ink display system using a Raspberry Pi Zero 2W, complete with a custom Flask-based remote image hosting pipeline.

• Languages: Python, R, SQL, Java
• Machine Learning: PyTorch, Scikit-learn, LangChain, HuggingFace, Pandas, NumPy
• Cloud & Infrastructure: Azure, GCP, Kubernetes, Docker, Kafka, gRPC, REST APIs
• Methods: Causal Inference, Time Series Forecasting, Bayesian Modeling, RAG, Computer Vision

• LinkedIn: linkedin.com/in/ishaan-agw
• Email: ia299@cornell.edu