Computer Vision Research Engineer

• May 2023 - Present

● Integrated different transformer-based text encoders in lora training of stable diffusion models ● Developed a background removal and replacement system based on stable diffusion models ● Trained several models based on textual inversion method for defect detection in images ● Devised a fast visualization and exploration system for large imagery databases based on similarity indexing ● Integrated different transformer-based text encoders in lora training of stable diffusion models ● Developed a background removal and replacement system based on stable diffusion models ● Trained several models based on textual inversion method for defect detection in images ● Devised a fast visualization and exploration system for large imagery databases based on similarity indexing

Senior Computer Vision Engineer

• Aug 2022 - Jul 2023

● Created a real-time face segmentation module with 20-30 ms of runtime on a normal Corei7 CPU and improved IoU to 98% and dice score to 97% on a dataset of 4k samples ● Established a skin segmentation module based on knowledge distillation, resulting in a significant improvement of up to 25% in both IoU and Dice Score ● Formed a real-time face anti-spoofing based on contrastive learning and FastViT with 97% accuracy and 15 ms of runtime on a common Corei7 CPU ● Implemented a real time face alignment module with the speed of 17ms, handling more than 40k samples of api requests per day ● Made a multi-class NSFW detection model, achieving 6% of accuracy more, and 50% less trainable parameters than the existing SOTA models ● Built a face authorization system, handling more than 40k successful cases daily and improved the precision and recall of the system by a margin of nearly 12% and 14% respectively ● Designed a BERT-based spell checker for the Persian language, achieving WER of lower of 3.5% on a dataset of 20 million Persian sentences, outperforming the SOTA methods in the literature ● Developed a CCTV-based self-supervised person reidentification model with Top-1 accuracy of more than 87% ● Mentored and coached 4 other teammate for Deep learning and Pyotrch Show less

Machine Learning Engineer

• Apr 2021 - Feb 2023

Missions: - Project 1 ● Developed a cross-platform software using PyQt for a private hospital for recording medical data from Traumatic brain injury (TBI) ● Equipped with features for training and evaluating different supervised machine learning-algorithms with a user-friendly UI ● Implementing SQL database and the required connections ● Testing and Maintaining - Project 2 ● Designed a document reader for a law firm for automating customer services ● Paragraph segmentation, signature extraction, OCR - Project 3 ● Implemented a user-friendly GUI-based app for cataract detection for a private medical services center ● Utilized light-weight deep models for better performance with low cost ● Achieved a trustworthy performance of 98.87 TPR on local dataset Show less

Machine Learning Engineer, Computer Vision Researcher

• Feb 2021 - Jul 2022

● Collected a dataset comprising more than 10k samples and made a semi-automated pipeline for labeling required for localization and recognition and enhanced the overall CER by integrating self-attention modules. ● Designed a content-based image retrieval system based on a self-supervised approach, achieving the 99% of human voted relevancy factor, while decreasing the pipeline's inference time to lower than 60ms. ● Implemented a novel seq2seq-based soccer match summarizer system capable of finding the most critical scenes, with more than 91% of exact frame match rate. ● Created a face retrieval system in movies achieving more than 97% precision and 95% recall for 1000 Iranian celebrities ● Built sound preprocessing modules, including human voice detector, sound diarization, and sound noise removal ● Training and maintaining Wav2vec and DeepSpeech for the Persian language, attaining 1.43 WER on “common voice” benchmark dataset ● Coached and pair-programmed for 3 newcomers in front and backend technologies over a course of 3 month Show less

Research Assistant

• Feb 2020 - Jul 2022

Software Developer

• Jun 2019 - May 2020

As a programmer who studied Civil Engineer in his Bachelor degree, I have the ability to do cross works between the field of engineering and computer. Thus, in Azaran Industrial Structures, I was responsible for multiple programs which were specifically designed for facilitating engineering procedures. It was a great opportunity because it gave the opportunity to widen my horizon in many hot areas within Civil Engineering, BIM and Augmented Reality-Aided Engineering protocols. As a programmer who studied Civil Engineer in his Bachelor degree, I have the ability to do cross works between the field of engineering and computer. Thus, in Azaran Industrial Structures, I was responsible for multiple programs which were specifically designed for facilitating engineering procedures. It was a great opportunity because it gave the opportunity to widen my horizon in many hot areas within Civil Engineering, BIM and Augmented Reality-Aided Engineering protocols.

Researcher

• Feb 2012 - Dec 2014

My coding skill and my knowledge in the field of Artificial Intelligence made a good choice for working in the cross domains between Artificial Intelligence and Civil Engineering. As a bachelor student, in the lab, I have researched on the analyzing the creep behavior of concrete beams and seismic behavior of steel structures with machine learning. This experience empowered my knowledge in scientific data collection, data pre-processing, applying various classical machine learning algorithms and deep learning based algorithms. Also, working under my supervisor's supervision, I researched on damage analysis of wind turbines using computer vision techniques. Show less