About Roshan
Roshan is currently a PhD Candidate in the Construction Engineering and Management program
at Oregon State University, and a data scientist intern at Skanska USA Building Inc. He
received his second Master's degree in computer science from Oregon State University.
Programming Skills
-
Python
90% -
C++
60% -
HTML
60% -
MATLAB
70% -
R
50%
Frameworks
-
Pytorch
70% -
Scikit-Learn
70% -
OpenCV
70% -
NLTK
60% -
Spacy
40% -
Flask
40%
Tools
Power-BI
Tableau
Jupyter
Pandas
Seaborn
SQL
GitHub
VS Code
Machine Learning
Deep Learning
Computer Vision
Soft Skills
-
Self Starter
Strategic Thinker
Leads Projects from Start to Finish
Resume
Professional Experience
-
Data Science Intern
Sep 2022 - Present
• At Skanska, I get involved in the daily tasks/meetings of project engineers, schedulers, safety inspectors, superintends, innovation team, and data analysts to learn and identify opportunities for scalable automation:
• Proposed idea, developed, and deployed a web application to reduce RoI. The web application aids design reviewers prevent design issues creep into construction phase. 20k RFI text records were cleaned, preprocessed, visualized, EDA performed using Power-BI, mySQL, Pandas. NLP, Computer Vision, Chat-GPT API used to match RFIs with drawings and summarize results. Flask used to develop the web portal. Product to be optimized, debugged, hosted on Skanksa Server. User experience to be collected. Effectiveness of product to be evaluated, potentially i mproved in Q2, and Q3.
• Answered questions related to project delay predictability using milestone delays. Cleaned, merged, preprocessed, binned, feature engineered, visualized, and tested correlation of schedule data from 256 projects using Seaborn, Scikit-Learn. Used ensemble of ML algorithms to predict delays.
• Proposed idea to automate out portion of scheduler staff time by automatically predicting the cause of delay using the log of weekly meetings with superintendent. EDA done on 1k+ records of comments from field crew, comments classified into 11 categories for delay using statistical tests, bag of words, and machine learning.
Education History
-
Master Of Engineering in Computer Science
2021 - 2022GPA: 3.83
Courses: Machine Learning, Computer Vision, Methods of Data Analysis I & II, Scalable Querying & ML, Database Management Systems, Parallel Programming, Computer Graphics Intro/Shaders, Analysis of Algorithms, Operating Systems II, Theory of Computation, Programming Languages -
Master Of Engineering in Computer Science
2018GPA: 3.83
Courses: Machine Learning, Computer Vision, Methods of Data Analysis I & II, Scalable Querying & ML, Database Management Systems, Parallel Programming, Computer Graphics Intro/Shaders, Analysis of Algorithms, Operating Systems II, Theory of Computation, Programming Languages
Research Publications
- Roshan Panahi; Joseph Louis; Nicholas Aziere; Ankur Podder, and Colby Swanson, Identifying Modular Construction Worker Tasks Using Computer Vision, i3CE, 2021. 'https://doi.org/10.1061/9780784483893.118'
- Roshan Panahi; Joseph Louis; Nicholas Aziere; Ankur Podder, and Colby Swanson,Tracking Volumetric Units in Modular Factories for Automated Progress Monitoring Using Computer Vision, CRC, 2022 'https://doi.org/10.1061/9780784483961.086'
- Roshan Panahi; Joseph Louis; Ankur Podder, and Colby Swanson, Automated Progress Monitoring and Bottleneck Detection in Modular Construction Factories Using Computer Vision, Sensors, 2023 'Forthcoming'
- Roshan Panahi; Joseph Louis; Ankur Podder, and Colby Swanson, Automated Assembly Progress Monitoring in Modular Construction Factories Using Computer Vision-based Instance Segmentation, i3CE 2023, 'Forthcoming'
- Roshan Panahi; Joseph Louis; Ankur Podder, and Colby Swanson,Automated Progress Monitoring in Modular Construction Factories Using Computer Vision and Building Information Modeling, ISARC, 2023 'Forthcoming'
- Roshan Panahi; Joseph Louis; Ankur Podder, and Colby Swanson,Request for Information (RFI) Recommender System for Pre-Construction Design Review Application Using Natural Language Processing, Chat-GPT, and Computer Vision , i3CE 2023, 'Forthcoming'
Portfolio
Contact Informations
- E-mail: panahir@oregonstate.edu