Sepehr Ghafari
Technical Product Owner at bimodel- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
Persian Native or bilingual proficiency
-
Turkish Native or bilingual proficiency
-
English Professional working proficiency
Topline Score
Bio
Credentials
-
Project Initiation: Starting a Successful Project
CourseraAug, 2023- Nov, 2024 -
Foundations of Project Management
CourseraApr, 2022- Nov, 2024 -
Machine Learning
CourseraFeb, 2022- Nov, 2024 -
ICCKE 2020 Reviewer Member
Ferdowsi University of MashhadOct, 2020- Nov, 2024
Experience
-
bimodel
-
France
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Technical Product Owner
-
Jun 2022 - Present
Upon joining the company, a simultaneous goal was initiated to launch the Modelic web application in 2022. The objective of this application is to establish a B2C platform that offers diverse BIM models to customers on both small and large scales. Through this platform, customers, company administrators, and engineers can collectively navigate all project stages—ranging from initial consultations, pricing, and project execution by the company's architect engineers, to monitoring project progress and facilitating payments—efficiently and collaboratively. The integration of artificial intelligence in segmenting architectural models further accelerates project progress and reduces architects' workload. In this project, my primary responsibility involves constructing the backend utilizing the FastAPI framework. My duties throughout this project encompass: ● Developing the backend using the FastAPI framework in conjunction with a MongoDB database. ● Setting up the project's initial CI/CD and DevOps infrastructure utilizing NGINX, Docker, Docker Compose, and GitHub. ● Developing an object storage system using Minio on server. This system facilitates the uploading and downloading of architectural data by both clients and architects. ● Collaborating closely with the development team to ensure that the project's technical facets harmonize with the overarching product vision. ● Crafting user stories and effectively managing the product backlog. ● Maintaining close collaboration with developers and overseeing task management. ● Collaborating with stakeholders on app goals and user interface design, while taking into financial and time constraints. ● Developing a segmentation model using TensorFlow's Keras library and the WUNET architecture. (At R&D phase) Show less
-
-
-
National Geographical Organization
-
Tehran, Iran
-
Artificial Intelligence Engineer
-
Mar 2021 - Jun 2022
I submitted a proposal for military service and was accepted into the "elite military" plan. During that period, I was assigned two projects: 1. Production of Land Cover Map using Sentinel 2 Satellite Images and Google Earth Engine Processing Platform. The primary objective of this project was to create a precise land cover map. This involved applying advanced machine learning techniques such as supervised and unsupervised methods (k-means, G-means, artificial neural networks, support vector machines, random forest, multi-label classification), along with utilizing Sentinel 2 satellite data, digital elevation models, and additional data sources. Furthermore, this project entailed precise result analysis and validation. 2. Construction of a Highly Accurate and Efficient CNN Model for Building Change Detection. For this project, my goals included: ● Developing a robust building detection model using UNET segmentation models with a targeted accuracy of over 90%. ● Developing a dependable model for change detection in buildings, maintaining an accuracy level of above 90%. ● Generating a DSM (Digital Surface Model) map by utilizing stereo images acquired from an aircraft, and subsequently employing it as supplementary data for model training. ● Identifying the optimal parameters for effective model training. ● Thoroughly examining the results to improve data quality and model performance, including finding any issues with the available data and labels. ● Performing preprocessing on the primary images to rectify any existing data imperfections. Show less
-
-
-
Ferdowsi University of Mashhad
-
Iran
-
Higher Education
-
700 & Above Employee
-
Research Assistant Professor
-
Sep 2017 - Feb 2021
During my academic journey, I developed a profound interest in artificial intelligence, neural networks, and image processing. My fascination extended to leveraging practical and extensive datasets within the industry, primarily utilizing Python. From my very first year, I embarked on simultaneous efforts with esteemed university professors to craft two articles. These collaborative endeavors culminated in successful publications. Consequently, I was approached to share my knowledge through providing lectures and guidance to fellow students in the computer department. Some of the topics I had the privilege of teaching and advising on included: ● Introduction to Python ● Essential Python libraries: numpy, pandas, matplotlib, opencv, keras, and more ● Working with satellite and aerial images, encompassing the analysis of distinctive features for each type ● Machine learning algorithms and the utilization of libraries such as scikit-learn ● Training and testing neural networks while rigorously scrutinizing the ensuing outcomes ● In-depth exploration of Convolutional Neural Networks (CNNs) ● Comprehensive guidance on composing and submitting articles to be included in ISI-indexed journals Show less
-
-
-
Farazamin Consulting Engineers
-
Tehran, Iran
-
Survey Engineer
-
Sep 2015 - Sep 2016
My responsibilities encompassed utilizing AutoCAD Civil 3D software for precise drawings and updating the map of Tehran. My responsibilities encompassed utilizing AutoCAD Civil 3D software for precise drawings and updating the map of Tehran.
-
-
Education
-
2017 - 2020
Ferdowsi University of Mashhad
Master's degree, Remote sensing, GIS -
2011 - 2016
Islamic Azad University,Science And Research Branch
Bachelor's degree, Surveying Engineering
Community
