Looking for a job

Sarvenaz Memarzadeh

@sarvenazmemarzadeh

Message

I build end-to-end machine learning, forecasting, recommendation, and LLM analytics systems.

United States

Message

At Allstate Health Solutions Products, I build end-to-end machine learning systems for forecasting, recommendations, and analytics. My time-series forecasting system achieved over 90% prediction accuracy, while my LLM-driven pipelines turn structured data into natural-language insights in Power BI dashboards.

Previously at Illumina, I built unsupervised learning pipelines to cluster millions of sequencing basecalls and identify signal-quality patterns. I also introduced a covariance-based signal-to-noise metric and monitoring tool to detect quality degradation across sequencing cycles.

My work combines Python, SQL, Azure, MLflow, LangChain, scikit-learn, and PyTorch with a Ph.D. in Electrical Engineering from the University of Maryland. I enjoy translating complex signals and analytics into practical, decision-ready products.

Experience

Work history, roles, and key accomplishments

Current

Senior Data Scientist

Current

Allstate

Jul 2025 - Present (1 year 2 months)

Built an end-to-end time-series forecasting system using advanced ML models, achieving over 90% prediction accuracy. Developed a recommendation system and an LLM-driven summarization pipeline on Microsoft Azure, and led the development of interactive PowerBI dashboards.

Time series Forecasting Machine Learning Recommendation Systems LLM Azure Power BI Python SQL

Senior Data Scientist

Illumina, Inc

Sep 2021 - Jul 2025 (3 years 10 months)

Built an unsupervised machine learning pipeline using DBSCAN and Gaussian Mixture Models to cluster sequencing basecalls. Designed a novel Signal-to-Noise Ratio metric and developed a time-series monitoring tool for early anomaly detection.

Unsupervised Learning DBSCAN Gaussian Mixture Model Time Series Analysis Signal Processing Python

Graduate Research & Teaching Assistant

University of Maryland - College Park

Aug 2016 - May 2021 (4 years 9 months)

Developed neural network models to predict material optical properties and implemented unsupervised machine learning algorithms for signal processing. Mentored undergraduate students and was awarded the Maryland Energy Innovation Institute fellowship.

Neural Networks Unsupervised Learning Signal Processing Mentorship