About
Subhajit Das graduated with a Ph.D. in Computer Science from Georgia Institute of Technology, USA. His research interests include user modelling, interactive machine learning, and interactive model optimizations bringing end-users' expertise in the modelling pipeline. Over the years, he has extensively worked with domain experts to build interactive AI applications to broaden access to ML and data science. His work with Visual Analytics Lab at Georgia Tech, and with DARPA's explainable AI initiative explores, "ML model selection, interactive optimization, and model interpretability for data exploration and decision making." He has previously developed interactive AI applications with industry partners including Autodesk, BOSCH, GTRI, UPS, 3M, and Microsoft based on extensive user research.
Speciality
- In-depth expertise in full stack web development and software engineering with AI-based applications.
- Extensive industry and research experience with AI / Machine learning models, web frameworks, and interaction/visual design to include user feedback in modeling.
- Expertise in synthesizing personas and user needs by usability testing (in-person and remote).
- Experienced with user data collection, utilising methods such as contextual inquiry, usability testing, heuristic evaluation, interviews, longitudinal design studies, and crowdsourced surveys.
- Proficient in writing and visual communication, demonstrated through high-quality publications in journals.
Researcher & Developer.
- Website: https://subhdas.github.io/
- City: Atlanta : Georgia, USA
- Degree: PhD Computer Science
- Email: das.connect@outlook.com
Working with domain experts, have developed novel interaction techniques with multiple machine learning models (Auto-ML) using interactive model selection, model steering, and interactive objective functions. Have conducted qualitative and quantitative user studies to validate tools and techniques as part of my research. Interests : Interactive machine learning, Model optimization, User Research, HCI + AI, Visual Analytics
Skills
ProgrammingResume
PDFSumary
Subhajit Das
Research : Working with domain experts, have developed novel interaction techniques with multiple machine learning models (Auto-ML) using interactive model selection, model steering, and interactive objective functions. Have conducted qualitative and quantitative user studies to validate tools and techniques as part of my research. Interests : Interactive machine learning, Model optimization, User Research, HCI + AI, Visual Analytics
- Atlanta, Georgia, USA
- Georgia Institute of Technology, College of Computing, USA
- das.connect@outlook.com
Education
PhD. in Computer Science - HCI and Applied Machine Learning
2014 - 2020 (expected)
Georgia Institute of Technology, Atlanta, USA
Researching novel interactions and interfaces to allow curation, optimization, selection and steering of multiple machine learning models to broaden access to ML. Working with domain experts, developed methods to interactively navigate model space, optimize and select models using techniques such as multi-model steering, interactive objective functions, and conflict resolution in objective specification. Collaborated with Tufts University, and University of Wisconsin to prototype ML applications to enable interactive data augmentation and model selection to democratize access to AI solutions.
Master of Science in Computer Science - Visual Analytics
2014 - 2015
Georgia Institute of Technology, Atlanta, USA
HCI, Visual Analytics, Machine Learning, A.I., Graphics and Optimization, UX Research, Experimental Design | Data Science Certificate from John Hopkins University (Coursera.org)
Master Architecture & Design Computing
2008 - 2009
University of Pennsylvania, Philadelphia, USA
Evolutionary Algorithms, Digital Fabrication, Swarm Intelligence, Form and Algorihtms, Multi Agent Systems, Strategy Management (PgDBA) | Leed AP
Professional Experience
UPS Industrial Engg. | Visualization Intern
May 2020 - Aug 2020
Atlanta, GA, USA- Developing a geo-visualization system to make sense of origin-destination based optimization models. (NDA)
Bosch Research | Visual Analytics Intern
May 2019 - Aug 2019
Sunnyvale, CA, USA- Interpretable machine learning and explainable AI using matrix optimizations on multiple data formats. Crowd-sourced user feedback from Amazon’s Mechanical Turk. (NDA)
Autodesk | Software Engineering Intern
May 2018 - Aug 2018
Boston, MA, USA- Instantiated microservices to unify data and parameters across multiple desktop applications under one umbrella data schema.
- Prototyped a use case of collaborative shared workspace using common data standards.
Microsoft Research | Research Intern
May 2017 - Aug 2017
Cambridge, UK- Prototyped STORYBOARD - a web-based slideshow builder, facilitating users to focus on content development by automating slide organization, management and presentation of slide decks.
Autodesk with Perkins Will | Software Intern
Jan 2016 - Aug 2016
Atlanta, GA, USA- Prototyped a plugin written in C# generating a large number of automated 3D floor plans helping healthcare planners pick a suitable option.
- Conducted user study with experts from Perkins Will to further validate the tool.
Autodesk | Software Intern
Jun 2015 - Aug 2015
Boston, MA, USA- Worked on the development of a web-based 3D design platform in ThreeJS.
- Incorporating domain expertise prototyped an extension written in Python enabling automated delivery of 3D model options for facility planning.
Publications
Journal/Conference Papers
-
QUESTO: Interactive Construction of Objective Functions for Visual Analytics, Computer Graphics Forum, EuroVis 2020, Subhajit Das, S. Xu, M. Gleicher, R. Chang, A. Endert
-
Gaggle: Visual Analytics for Model Space Navigation, Graphics Interface Conference 2020, Subhajit Das, D. Cashman, R. Chang, A. Endert
-
Geono-Cluster: Interactive Visual Cluster Analysis for Biologists IEEE-Transactions on Visualization and Graphics 2020 Subhajit Das, B. Saket, B. C. Kwon, A. Endert
-
BEAMES: Interactive Multi-Model Steering, Selection, and Inspection for Regression Tasks, IEEE Computer Graphics and Applications 2019 Subhajit Das, D. Cashman, R. Chang, A. Endert [ BEST PAPER AWARD ]
-
A User-based Visual Analytics Workflow for Exploratory Model Analysis, Computer Graphics Forum, EuroVis 2019 D. Cashman, S. R. Humayoun, F. Heimerl, K. Park, Subhajit. Das, J. R. Thompson, B. Saket, A. Mosca, J. Stasko, A. Endert, M. Gleicher, R. Chang
-
Comparing Apples and Oranges: Taxonomy and Design of Pairwise Comparisons within Tabular Data, CHI Conference on Human Factors in Computing 2019, P. M. Law, Subhajit Das, R. Basole
-
Interpreting non-flat surfaces for walkability analysis, SIMAUD 2019, M. Shwartz, Subhajit Das
-
Podium: Ranking Data Using Mixed-Initiative Visual Analytics, IEEE Transaction on Visualization and Graphics 2018, E. Wall, Subhajit Das, R. Chawla, A. Endert
-
Energy Model Machine (EMM) - Instant Building Energy Prediction using Machine Learning, eCAADe Conference 2017, R. A. Mohammad, Subhajit Das, T. Barry, M. Ian and H. Anthony
-
Space Plan Generator: Rapid generation & evaluation of floor plan design options to inform decision-making, Acadia Conference 2016, Subhajit Das, C. Day, J. Haymaker, A. Hauck, D. Davis
Workshop Papers
-
InMacs: Interactive modeling and comparison of sentiments from sequence data, Workshop on Data Science with Human in the Loop (DaSH), at KDD 2020 conference 2020, Subhajit Das, F. Dutt
-
BEAMES: Interactive Multi-Model Steering, Selection, and Inspection for Regression Tasks, Visual Data Science Symposium at IEEE VIS Conference 2018 , Subhajit Das, D. Cashman, R. Chang, A. Endert [ BEST PAPER AWARD ]
-
ISPARK: Interactive Visual Analytics for Fire Incidents and Station Placement, Workshop on Interactive Data Exploration and Analytics (IDEA) at KDD conference 2015, Subhajit Das, A. McCarter, J. Minieri, N. Damaraju, S. Padmanabhan, D.Horng (Polo) Chau
Posters
-
Podium: Ranking Data Using Mixed-Initiative Visual Analytics, CRA-W Grad Cohort, San Diego, CA, April 2016, E. Wall, Subhajit Das, R. Chawla, and A. Endert
-
Mitigating fire risks using visual and data analysis, Bloomberg: Data for Good Exchange, New York, USA, Sep 2015, N. Damaraju, Subhajit Das, A. McCarter, J. Minieri, S. Padmanabhan, D. H.(Polo) Chau
-
Design ornamentation & fabrication by multi agent systems, SIGGRAPH Posters 2012, Subhajit Das, F. Dutt
Contact
Please connect to talk ideas !
Location:
Atlanta, Georgia, USA
Email:
das.connect@outlook.com