Fereshteh Soltani Email & Phone Number

Los Angeles, Ca, Us

Santa Mônica , Califórnia , Us

New York, New York, Us

San Jose, California, Us

▪Implement a semantic text similarity system with deep learning techniques, semantic entity analysis, and topic modeling.▪ Build word/phrase semantic embeddings, powering semantic search, and abstractive summarization.▪ Build an online learning algorithm to predict and obtain the desired outputs for new data points.▪ Build pipelines with distributed, large scale data processing platforms.▪ Deliver data-driven solutions using deep expertise to provide viable solutions, analysis, rapid iterations, and actionable insights.▪ Understand and align with business objectives, adjusting and planning accordingly for a rapidly changing deliverable work-flow.▪ Collect and synthesize requirements to create an effective feature roadmap.▪ Build out features for data collection and recommendation engines.▪ Mentor and lead junior scientists.

Analyze precipitation and evaporation (time series) datasets and develop a model to predict groundwater level fluctuations through the use of visualization, and mining techniques. -Visualize water budget time series data using Tableau and prepare user stories to create compelling dashboards to deliver actionable insights.-Efficiently preprocess (Energy, water) consumption datasets, including integration, transformation, cleaning and filtering to ensure accurate results for projects. -Design, develop, test a website application to show geo map with water information details, and developed a dynamic statewide water budget simulation applying Agile and using Powersim SDK, Microsoft Visual Studio, ASP.net, C#-Acquire data from primary or secondary Spatial data (PRISM Climate Group dataset) and maintain New Mexico statewide water assessment databases. (ArcSDE, SQL Server)

- Prepared Persian textual database consisting of letters with their corresponding pronunciations for Persian text-to-speech engine and text mining project.- Applied machine learning/text processing algorithms to extract phrases and name entities from Persian text.

Trento, Tn, It

- Utilized Phrase Detectives data to train anaphora resolution algorithms with Bart.- Interpreted data, analyzed results using statistical techniques and provided ongoing reports to stockholders.

- Developed facial emotion detection application using principal component analysis techniques.