• Hi! I'm
    Mohammad.

    Someone who is passionate about AI and programming, with a strong foundation in mathematics and exceptional problem-solving skills.

About

Who Am I?

Hi, I'm Mohammad Farsi, a Ph.D. graduate from the Communication Systems group at Chalmers University of Technology, one of Sweden's leading research institutions.

🎓 I am passionate about learning, applying, teaching, and sharing knowledge.

🔬 My research interests center around the application of machine learning and deep learning in autonomous driving systems, where I'm passionate about developing intelligent solutions for next-generation vehicles.
I also specialize in classical digital signal processing and advanced sensor fusion techniques, bridging the gap between traditional signal processing methods and cutting-edge AI technologies.

💻 I enjoy developing innovative solutions and contributing to open-source projects that advance our understanding of communication technologies and their real-world applications.

Machine Learning & AI Expert

Telecommunications Expert

Signal Processing Expert

My Work

Passion projects.

View CV

Sensor Fusion and Localization (Zenseact)

Advanced multi-sensor fusion for autonomous vehicle localization

Fusing multiple sensors including GNSS, camera, lidar, IMU to localize the car on HD and SD maps. Many methods are used including complex adaptive Kalman filters, hidden markov models, and other advanced probabilistic techniques for robust and accurate positioning in challenging environments.

GNN-based SD-Map encoder

Using GNN to encode SD-map and fuse it to Zenseact world view vector via cross attention

Advanced GNN architectures including GCN, GAT, VectorNet, and GraphSAGE to encode standard definition maps. Integrated with Zenseact's world view vector via cross-attention for enhanced autonomous vehicle localization.

Machine Learning (Chalmers)

Using ML for sensing the environment with optical fiber cables.

Optical fibers not just for transmitting data but as incredible sensors too. These slender strands of glass are now equipped to detect changes in their surroundings, providing us with invaluable information about the environment. It's a groundbreaking application that allows us to harness the power of light to sense and understand our world in remarkable ways..

Fiber Capacity (Chalmers)

Utilizing advanced mathmatical tools to derive bounds on the capacity of high-dimensional fiber optic channels. This prject required:
- Deep understaning of probability theory and directional statistics.
- Understaning and deploying convex and non-convex optimization algorithms.
- Implementing high-dimensional entropy estimators. .

DSP Developer (Faraz Telecom)

Implementing reciever digital signal processing chain.

Delving into receiver DSP chain, including timing synchronization, CFO compensation, channel estimation, equalization, decoding, and detection. I tackled hardware challenges such as phase noise and amplifier distortion, working with antennas, power amplifiers, and circuitry.

DSP Developer (Faraz Telecom)

Implementing reciever digital signal processing chain.

- Proficient in object-oriented programming languages like C++, Python, and Matlab.
- Intermidiate in Java.
- Intermidiate in Haskell (funcional programming).
Experience

Work Experience

Zenseact

AI/ML Engineer

Zenseact AB

2024 - Current
  • Developed advanced sensor fusion algorithms for autonomous vehicle localization using multi-modal data (GNSS, camera, lidar, IMU).
  • Implemented Graph Neural Network architectures (GCN, GAT, VectorNet, GraphSAGE) for SD-map encoding and integration with vehicle perception systems.
  • Created end-to-end localization and trajectory planning systems using Temporal Graph Networks and Spatio-Temporal GCNs.
  • Applied advanced probabilistic techniques including adaptive Kalman filters and hidden Markov models for robust positioning.
  • Developed high-performance algorithms in both Python and C++ for real-time autonomous driving applications.
Chalmers

Ph.D. Student

Chalmers University of Technology

Jan 2020 - June 2024
  • Deploying machine learning algorithms to determine the location of earthquakes using optical fibers.
  • Characterizing the full-dimensional capacity of a fiber using information theory tools.
  • Designed and developed DSP algorithms for phase, polarization, and polarization dependent loss compensation.
Chalmers

Software & DSP Developer

Faraz Telecom

Sep 2016 - Dec 2019
- Expertise in advanced radio algorithms, including:
  • Receiver DSP Chain: timing synchronization, CFO compensation, channel estimation/equalization, decoding, detection.
  • Hardware impairment (phase-noise, amplifier distortion, etc.) compensation
  • Hardware including antennas, power amplifiers, and circuitry.
- Development and simulations utilizing object-oriented programming in C++, Python, and Matlab
- Practiced agile project management as project leader in two projects.
- Collaborated in implementing algorithms on hardware (FPGA).
What I do?

Problems I like solving.

Machine Learning

Utilizing machine/deep learning algorithms to solve various problems. For instance:
- Data Analytics.
- Resource Allocation.
- Network Planning.
- Security and Fraud Detection.

Communication Engineering

- Develop, simulate, and analyze performance of DSP algorithms.
- Analyze future needs and requirements. - Optimizing resouces.

Research Towards Future technologies

- I would love to have an impact in research and development of future technologies.

Education

Education

Western University
University of Tehran 09/2012 - 09/2016 Bachelor of Science (BSc)
  • Major in Electerical Engineering
Western University
Tarbiat Modarres University 09/2016 - 02/2019 Master of Science (BSc)
  • Major in Communication Systems
Western University
Chalmers University of Technology 01/2020 - 06/2024 (expected) Ph.D.
  • Major in Communication Systems
Get in Touch

Contact