Sajjad Karimi

Sajjad Karimi

Postdoctoral Fellow in Biomedical Informatics at Emory University | Expert in Statistical Signal Processing, Machine Learning, and Biomedical Signal Analysis

Welcome to Sajjad Karimi’s Academic Profile

About Me

I am a Postdoctoral Fellow in the Department of Biomedical Informatics at Emory University School of Medicine, specializing in biomedical signal processing, machine learning, and statistical modeling. My research focuses on advancing health diagnostics through the analysis and modeling of complex multi-channel biomedical signals, including brain and heart data.

My expertise spans statistical signal processing, machine learning, and advanced modeling techniques to extract biomarkers that improve cardiovascular and neurophysiological health assessments. I am passionate about mentoring students, promoting interdisciplinary collaboration, and publishing research that drives innovation in biomedical signal processing.

Research Interests

My experience includes processing & modeling biomedical signals & financial time series using the following areas:

  • Statistical Signal Processing
  • Machine Learning
  • Coupled Hidden Markov Models
  • Statistical Inference & Modeling
  • Mathematical Optimization
  • Biomedical Signal Analysis
  • Brain-Computer Interfaces
  • Cardiovascular Signal Processing

Current Research

Doctoral Research: “A New Coupled HMM Framework & Applications in Multi-channel Brain Signal Processing”

My research focuses on the Latent Structure Influence Model (LSIM), a type of Coupled Hidden Markov Model (CHMM). Our contributions are divided into theory and application parts:

Theory: We proposed a new fast approximate inference algorithm for LSIMs, proved the convergence of the proposed learning framework, and developed an EM algorithm for parameter re-estimation.

Applications: We proposed two new applications using LSIMs including channel fusion for improving sleep staging performance and a new estimator for brain connectivity. We found that LSIM had better performance than HMMs and CHMMs.

Professional Experience

Current Position

Postdoctoral Fellow | Emory University School of Medicine | October 2024 - Present

As a Postdoctoral Fellow in Biomedical Informatics, I contribute to advancing health diagnostics by analyzing and modeling complex multi-channel biomedical signals. My work leverages statistical signal processing, machine learning, and advanced modeling techniques to extract biomarkers that improve cardiovascular and neurophysiological health assessments.

Previous Experience

Senior Data Scientist & Quantitative Analyst | Charisma Financial Information Processing | April 2022 - September 2024

Developed and implemented an AI-driven portfolio management system using MLOps, machine learning models, MDP, and Dynamic Bayesian Networks. Applied statistical models to predict and analyze the options market, achieving a 10% excess return on a $30,000 test fund in four months. Optimized portfolios using the Black-Litterman model, successfully managing a $2.5 million fund with a $23 million turnover over two years.

Education

  • Ph.D. in Electrical EngineeringSharif University of Technology2015-2023
  • M.S. in Bioelectric EngineeringSharif University of Technology2012-2014GPA: 18.59/20 (Rank: 2/12)
  • B.S. in Electrical EngineeringShiraz University2008-2012GPA: 18.20/20 (Rank: 1/32)

Selected Publications

Karimi, S., & Shamsollahi, M. B. (2023). Tractable maximum likelihood estimation for latent structure influence models with applications to eeg & ecog processing. IEEE Transactions on Pattern Analysis and Machine Intelligence.

Karimi, S., & Shamsollahi, M. B. (2020). Tractable inference and observation likelihood evaluation in latent structure influence models. IEEE Transactions on Signal Processing, 68, 5736-5745.

Karimi, S., & Shamsollahi, M. B. (2022). A New Post-Processing Method Using Latent Structure Influence Models for Channel Fusion in Automatic Sleep Staging. IEEE Journal of Biomedical and Health Informatics.

For a complete list of publications, please visit my Publications page or Google Scholar profile.

Contact