I love deep learning. My whole interest in science is about AI and practicing and researching the edge-tech fields in deep learning. I have professional experience more than 5 years. I have earned knowledge in CV, NLP, and Speech and keep updating myself with recent papers and breakthroughs in this field. I have experience in a startup environment and had experience in the entrepreneurship field. I have experience in computer vision, machine vision, image processing, GANs, transformers, ASR, NLP, and … I self-studied the whole Python and machine learning algorithms and programming skills by practicing and working. I am also trying to learn more about data science and MLOPS and am willing to learn and work on researchable projects. I enjoy every day of my life studying this field.

• We provide AI services and products. Such as Bread Counter.

• I earned experience by practicing machine vision and using computer vision in different fields of industrial monitoring, optimization, and image processing.
• I was the AI tech lead of a team and we were providing machine vision services such as classic methods of image processing (feature extraction) like SIFT, BRISK, FREAK, SURF, or ORB. DNN computer vision algorithms such as Image classification, object detection (Yolo, SSD, Detectron2, Fast R-CNN and …), Image segmentation (Mask R-CNN, U-net, Deeplab, and …), anomaly detection, OCR, Face Recognition, Object tracking, Key-point estimation (Hand pose, Body pose Face-mesh) and etc.
• I was responsible for team leadership and teamwork. I learned so much from my colleagues and improved my teamwork spirit by implementing Agile and Scrum best practices for our team.

• I spent an internship in Symo and earned experience with computer vision and generative models for classification and object detection tasks with CNNs and GANs.
• I was responsible for the Research and Development of the company and implementing models from papers and fine-tuning computer vision models for our clothes datasets.
• I was also involved in collecting the datasets using BS4 for crawling the images and cleaning the datasets and annotating for training classifiers and object detection models.

• This code demonstrates the usage of different language models for question-answering (QA) and document retrieval tasks using Langchain. The script utilizes various language models, including OpenAI's GPT, Google's Palm, and similarity search with Llama_index's retrieval approach, to provide answers to user queries based on the provided documents.

• Pytorch implementation of the CRNN model. In this repository, I explain how to train a license plate-recognition model with custom dataset using pytorch-lightning.

• This repository contains the most frequently used deep learning models and functions. Deep_Utils is still under heavy development, so take into consideration that many features may change in the future.

• This is our work with x-ray chest covid datasets. Using active learning we managed to build a classifier with 99% accuracy to detect covid symptoms in images. In both Keras and PyTorch frameworks.

• This algorithm is based on YOLO object detection and will count every bread dough that will enter the oven of the bakery. The purpose of this software is to check whether the bakery will bake the correct amount of bread that the government gave them flour based on that amount.

• This algorithm is based on MediaPipe API and we developed the datasets by adding more images of hands it can be used for hand tracking and even also will be used for object detection integration.

• This algorithm is based on MediaPipe API and we developed the simple model into a model with more accuracy it can be used for body tracking and even also will be used for object detection integration including a person's body.

• This tool has multiple forms of paraphrasing modes and different grammar in English for paraphrasing. Repo

• This time I used the new Yolo V7 version to train a model for medical datasets in Roboflow and get the inference with 2.8 ms.

• In this repository, I trained this dataset with both Pytorch and Keras frameworks. Due to the imbalanced and corrupted datasets, the convergence was not good enough. But in this repository, I implemented transformers in vision with resnet18 backbone in Pytorch for research purposes. Fer Emotion Recognition using VIT

Email: [email protected]
LinkedIn: https://www.linkedin.com/in/vargha-khallokhi
Phone Number: 09396463632