PH monogram
Pejman Habashi Staff Software Engineer

Research & Development

Building practical ML, CV, and 3D systems that hold up in production.

I work on perception pipelines, 3D reconstruction, and graphics-intensive software, translating research into reliable, maintainable products.

Download CV Connect on LinkedIn
PH
  • Current Role Software Engineer Staff, Qualcomm (2023–Present)
  • Focus Areas Machine Learning / Computer Vision / 3D Reconstruction / Graphics
  • Publications DBLP · Google Scholar
  • Profiles GitHub · GitHub (Tessonics)

About

I am an R&D software developer focused on machine learning, computer vision, 3D reconstruction, and graphics. My work centers on building software that is research-informed but production-ready.

Experience

Software Engineer Staff Qualcomm
2023–Present
  • Develop deep neural networks for environment perception.
  • Convert multi-camera inputs into bird's-eye-view (BEV) representations.
  • Support perception systems across Linux and QNX targets.

Tools: C++, OpenCV, Python, PyTorch, OpenMMLab, Linux, QNX, Docker, Argo, JIRA.

R&D Software Developer Tessonics Inc.
2020–2023
  • Implemented online 3D volume rendering of captured data.
  • Built oblique slicing and measurement tools for volumetric data.
  • Registered ultrasound surfaces to CT using an ICP-based approach.
  • Ran DNN inference in C++ for TensorFlow and torch models.

Tools: C++, OpenGL, OpenCV, Python, libTorch, TensorFlow, U-Net, dearImGui, GLSL.

R&D Software Developer AIS Tech Group
2017–2020
  • Designed and implemented industrial machine vision systems.
  • Built visual targets supporting over one million unique IDs.
  • Developed 3D pose math libraries for geometry calculations.
  • Implemented deep-learning-based human tracking.

Tools: C#, C++, OpenCV, Python, YOLO, SSD, PyTorch, TensorFlow.

Machine Vision Researcher Radix Inc.
2016–2017
  • Verified camera calibration pipelines and improved accuracy.
  • Implemented calibration upgrades for production systems.

Tools: Python, HALCON, OpenCV, C#, C++.

Senior Java Developer iNextWeb
2013–2015
  • Optimized next-generation web search engines.
  • Improved performance by distributing MongoDB across nodes.

Tools: Java, MongoDB.

Software Developer Maharan Engineering Co.
2006–2009
  • Developed telecom systems using distributed networking for automated train signaling.
  • Built GUIs to display train locations and control railroad systems.

Tools: C/C++, C#, SCTP.

Research Assistant

University of Windsor · 2013–2018

Human activity recognition, feature engineering, and classical ML pipelines.

Colloquium Representative

University of Windsor · 2016–2017

Coordinated weekly computer science colloquium meetings.

Research Assistant

Sharif University of Technology · 2006–2008

Created a practical Persian speech synthesis system using unit selection.

Education

PhD, Computer Science

University of Windsor · 2018 · Windsor, Canada

Human Activity Recognition · Multimedia Research Lab

MS, Artificial Intelligence

Sharif University of Technology · 2008 · Tehran, Iran

Speech synthesis by unit selection · Speech Processing Lab

BS, Computer Engineering

Sadjad University of Technology · 2006 · Mashhad, Iran

Computer engineering fundamentals and systems design.

Honors

Best Presentation

CSCON · University of Windsor · 2015 (3rd place)

Scholarship Recipient (GSS)

University of Windsor · 2015, 2017

Second Best GPA

Sadjad University of Technology · 2006

Best GPA (Semesters)

Sadjad University of Technology · Spring/Fall 2004, Fall 2005

Semi-finalist

National Mathematics Olympiad · Mashhad, Iran · 2002

Semi-finalist

National Computer Olympiad · Mashhad, Iran · 2001

Skills

Core Languages

C/C++, Python, C#, Java, Go, Bash

ML & Vision

Deep Learning, Machine Vision, Camera Calibration, Stereo Vision

Libraries & Tools

OpenCV, PyTorch, TensorFlow, OpenGL, HALCON, dearImGui, NumPy

Systems & Platforms

Linux, QNX, Docker, Windows, macOS, CI/CD, Multithreading

Selected Publications

  1. Disparity-Augmented Trajectories for Human Activity Recognition (2021). Evolutionary Intelligence.
  2. A Better Trajectory Shape Descriptor for Human Activity Recognition (2017). International Conference on Image Analysis and Recognition (ICIAR), pp. 330–337.
  3. Trajectory-based human activity recognition from videos (2017). Advanced Technologies for Signal and Image Processing, IEEE.
  4. The bag of micro-movements for human activity recognition (2015). In Image Analysis and Recognition, Springer, pp. 269–276.
  5. Automatic license plate recognition: A comparative study (2015). IEEE ISSPIT, pp. 635–640.
  6. A fast IFSR-based BIST approach for testing of digital integrated circuits (2012). IEEE ICCAS.
  7. Unit selection method for Persian speech synthesis using fast vox (2009). CSICC 2009.

Full list available on DBLP and Google Scholar.

Hobby Projects

Wordle Solver (2022)

Information theory-based solver for Wordle in any language.

Tech: C++, Go

Text Generation (2019)

Character-level LSTM for Persian text generation.

Tech: LSTM, PyTorch

Reversi Player (2013)

Minimax-based game AI that chooses optimal moves.

Tech: C++

Resources

Curriculum Vitae

Download the latest CV (PDF)

Publications

DBLP publication list · Google Scholar

Presentation

Image Processing slides (PDF)

Contact

For professional inquiries, please connect via LinkedIn or GitHub.