Sagun Karki

Data Scientist

B.S. in Computer Science and Data Science, Minor in Mathematics
University of Nebraska-Lincoln (Expected May 2026)

About Me

Iโ€™m a data scientist with a strong foundation in computer science, statistics, and applied mathematics. I am passionate about leveraging data to build intelligent applications and derive actionable insights. My core expertise lies in developing and deploying machine learning solutions, from data ingestion and processing to model training and inference.

Experience

Data Analyst

Nebraska Public Media

Jan 2025 โ€“ Present

Redesigned ETL pipelines with BigQuery, cutting runtime by 65% and enabling near real-time BI reporting for content and audience insights.

Data Science Intern

Raikes School (Allo Fiber)

Sep 2024 โ€“ May 2025

Developed an ensemble fraud detection system (HMM + SVM) for 40K+ IoT routers, reducing undetected fraudulent activities in SmartTown networks.

Machine Learning Intern

Nebraska Water Center

Feb 2024 โ€“ Jan 2025

Built a DNN for corn yield prediction using 20+ years of historical data, achieving 82% accuracy through feature engineering and PCA.

Projects

AI-Powered FAQ System ๐Ÿ”—

Python, Flask, FAISS, RAG

Led a team to build an AI assistant using a RAG pipeline with FAISS vector search for accurate, citation-backed answers.

Depression Detection

Python, Jupyter, XGBoost

Optimized an XGBoost classifier via hyperparameter tuning to achieve 94.5% accuracy in classifying signals from a Kaggle dataset.

Transit System Analysis

Python, Mapillary API, GraphQL

Engineered a data pipeline to process 30k+ images, correlating crash-risk signals with public transit data to map high-risk zones.

Technical Skills

Languages

Python, SQL, R, Bash, JavaScript

ML/AI & Data

Scikit-learn, TensorFlow, PyTorch, XGBoost, Hugging Face, FAISS, Pandas, NumPy, Power BI

Tools & Platforms

Git, GitHub Actions, Docker, Google Cloud (BigQuery, Dataform)

Python R JavaScript Bash SQL TensorFlow PyTorch Pandas NumPy Git GitHub Docker Google Cloud Linux

Publications

The US Economy as a Network ๐Ÿ”—

J Hawkins, S Karki

Comparison Across Economic and Environmental Metrics. Available at SSRN.

A dynamic approach for corn yield prediction to ensure agricultural resilience in the U.S. Midwest ๐Ÿ”—

A Mitra, S Karki, Et Al.

Neural networkโ€“based corn yield forecasting with adaptive integration of weather and historical yield data.

Better Safety Analyses through Smarter Data ๐Ÿ”—

M Elayan, S Karki, J Hawkins

Adding Open-Street-View and Traffic-Calibrated LBS Data to Pedestrian Crash Analysis in Lincoln, NE.