Welcome to my GitHub profile, where code, creativity, and research converge. 🚀🔬💡

I'm Borna Barahimi, a Machine Learning Researcher & Engineer with over four years of experience spanning academia and industry. Currently, I work as an Algorithm Designer at Cognitive Systems Corp., where I develop representation learning solutions for motion sequential data in WiFi sensing applications.

I recently graduated with an M.Sc. in Computer Science from York University (2024) with an A+ GPA, where I was a Research Assistant at NGWN Research Lab, supervised by Dr. Hina Tabassum. My research focused on WiFi sensing, self-supervised learning, and lightweight compression networks.

• Algorithm Designer (ML Research Intern) at Cognitive Systems Corp.

  • Working on unsupervised learning, quantization, and time-series analysis for WiFi sensing applications.
  • Developing a unified notion of sensing in WiFi-motion time-series data for IoT applications.
  • Implementing self-supervised learning methods like correlation reconstruction and VQ-VAE.

• Research Assistant at York University (2022 - 2024)

  • Led research on WiFi sensing using self-supervised learning and lightweight compression.
  • Developed a novel self-supervised framework for human activity recognition (HAR) with WiFi signals.
  • Designed a CSI compression model (RSCNet) for cloud-based sensing, reducing edge computation by 99%.
  • Peer-reviewed for top journals including NeurIPS, IEEE TCOM, IEEE Communications Letters.
  • Nominated for the Best Thesis Award at York University.

• Research Assistant at University of Tehran (2021 - 2022)

  • Worked on semi-supervised learning for Parkinson’s disease detection using online handwriting data.

• Fullstack Developer at VClinic (2019 - 2021)

  • Migrated a legacy healthcare system from Ruby on Rails to a modern microservice architecture.
  • Developed appointment scheduling, virtual visits, and electronic health records (EHR) solutions.

I specialize in representation learning frameworks that generalize across tasks without supervised signals. My work primarily focuses on:

• Self-supervised learning for low-labeled data and out-of-distribution detection.
• Machine learning applications for time-series analysis

• Context-Aware Predictive Coding: A self-supervised framework for WiFi sensing. Published in IEEE OJ-COMS and presented at NeurIPS 2024 Workshop on SSL. 📄 Paper | 💻 Code
• RSCNet: A dynamic CSI compression model for cloud-based WiFi sensing. Accepted at IEEE ICC 2024. 📄 Paper | 💻 Code

• 📧 Email
• 🌐 LinkedIn
• 📄 Google Scholar
• 🌍 Personal Website

🔔 Check out my repositories and give them a ⭐ if you find them useful!

Happy coding! 🎉👨‍💻🚀