Summary
Work History
Education
Skills
Research Experiences
Certification
Relevant Publications
Generic

SAJAD VAHEDIZADE

Saint Paul,MN

Summary

A diligent and passionate data science intern with strong organizational skills in statistical and analytical techniques to thrive in demanding digital intelligence processing environments. A resilient PhD candidate ready to build data-driven solutions to business problems with desire to exploit cutting-edge AI/ML frameworks (from data ingestion and preprocessing to model deployment and monitoring) in a full-time {\it data scientist} position.

Work History

Data Science Intern

3M | Consumer Business Group
Saint Paul, MN
06.2022 - Current
  • Collaborated on an invention submission focused on leveraging signal detection approaches in developing deep neural networks
  • Designed and implemented computer vision models including but not limited to: object detection, classification, and instance segmentation to meet product requirements in outcomes of interest
  • Compiled, cleaned, and manipulated datasets to evaluate model performance and deployed the models with ONNX format
  • Presented comprehensive progress reports to peers, and translated cost/benefits of ML product to executive stakeholders for business development
  • Gained hands-on experience on Azure, Databricks, and AWS cloud platforms, and developed database solutions in multiple SQL languages.

Education

Ph.D. - Civil Engineering – Earth And Atmospheric Science

University of Minnesota - Twin Cities
Minneapolis, MN
10.2022

Master of Science - Civil Engineering

University of Tehran
Tehran
2019

Bachelor of Science - Civil Engineering

Iran University of Science And Technology
Tehran
2016

Skills

    Technical:

  • Python (PyTorch, TensorFlow)
  • R
  • AWS (SageMaker, S3, EC2)
  • Data bricks (ML, SQL)
  • Microsoft Azure (Dev Ops, ML, SQL)
  • Analytical:

  • Computer Vision
  • Clustering and Classification
  • Statistical Analysis
  • Predictive Modeling
  • Deep Learning
  • Product Deployment

Research Experiences

I. Deep neural network for nowcasting and retrievals of global precipitation

  • Implemented an MLP deep neural network algorithm and a recurrent neural network (LSTM) to retrieve the precipitation occurrence and phase using 20 years of the IMERG dataset,
  • Increased the true positive rate up to 95%, while the false positive rate limited to around 4%.

II. Using signal detection to improve the snowfall retrievals in terms of type-I and -II errors

  • Designed a hypothesis testing problem to condition the precipitation retrievals through the Silhouette cluster analysis and the Neyman-Pearson hypothesis testing,
  • Reduced the error in the estimations and improved the accuracy by 6–8%.

III. Snowfall retrievals using Bayesian inference techniques

  • Developed a physically-informed Bayesian retrieval algorithm using the coincidences of CloudSat and GPM satellites to estimate the snowfall occurrence and its intensity with an RMSE of 0.15–0.2 mm/hr.

Certification

  • Machine Learning [CS]
  • Mathematics of Image and Data Analysis [MATH]
  • Nonlinear Optimization [EE]
  • Intro to Data Mining [CS]
  • Optimal Filtering and Estimation [EE]

Relevant Publications

1. S. Vahedizade, and Co-authors. On the Effects of Cloud Water Content on Passive Microwave Snowfall Retrievals. Journal of Remote Sensing of Environment, 2022. [IF: 13.85]
2. S. Vahedizade, and Co-authors. Passive Microwave Signatures and Retrieval of High-Latitude Snowfall Over Open Oceans and Sea Ice: Insights From Coincidences of GPM and CloudSat Satellites, IEEE Transactions on Geoscience and Remote Sensing, 2021. [IF: 8.12]
3. F.J. Turk, S. Vahedizade, and Co-authors. Applications of a CloudSat-TRMM and CloudSat-GPM Satellite Coincidence Dataset. Remote Sensing, 2021. [IF: 5.34]

SAJAD VAHEDIZADE