- Working on Computer Vision projects, researching, developing & deploying many services.

Mofid Securities is Iran’s largest private financial institution in the stock market exchange and the leading brokerage, serving over 15 million users with an annual income of over 500 million dollars. The company tackles diverse AI challenges to enhance financial services and customer experience.- Reduced model validation time by 50% and improved team productivity by 25% by developingand launching a backtester module for ML and DL evaluation.- Formulated and executed advanced ML & DL models to forecast movements in stock and gold markets, skillfully integrating macroeconomic factors, thereby enhancing predictive accuracy and financial strategies.- Achieved 90% validation accuracy in time‑series scenarios and reduced evaluation time by 50%by refining ML algorithms through t‑tests and A/B tests.

Mohaymen is a leading company in Iran, specializing in data infrastructure and big-data analytics, with a focus on projects related to taxes and financial transactions.- Reduced fraudulent transaction attempts by 40% and enhanced security for over 1,000 users bydesigning a Face Authentication engine with 99% accuracy in detection and 90% success in anti‑spoofing (ISO level 1).- Designed and implemented a comprehensive Face Recognition system; integrated modules like Aligning, Super‐resolution, and Head pose estimation into a system interacting with an Elasticsearch database, achieving 90% accuracy on a 1 million images dataset.- Advanced the Face Anti‑spoofing module with deep learning, enhancing dataset integrity for robustdetection of ISO 30107‑3 levels 1 and 2 attacks

Rasa Salamat is a health tech startup specializing in digital pathology, developing AI and image processing solutions to assist pathologists in cancer detection. Their technology is actively used byleading pathologists in Tehran’s most prominent hospitals.- Tripled image stitching speed by optimizing C++ algorithms with GPU acceleration, ensuringseamless image integration.- Halved upload times by optimizing image compression algorithms, boosting data handlingefficiency

A leading telecom solution provider in Iran, ranked among the top three in the industry.- Developed and implemented various ”Automatic Modulation Recognition (AMR)” techniques,including machine learning and deep learning models, improving signal modulation detectionaccuracy by 20%.

Worked on the "Football Video Summarization" project under the supervision of Dr. Mohammad Ali Akhaee. It was a research project and led to an MVP.- Engineered deep learning perspective transformation techniques, boosting player trackingaccuracy by 30% and reducing data processing time by 40%, aiding coaches and reporters indecision‑making - Some of the content of this project is attached.