I am currently pursuing a PhD in Human Computer Interaction (HCI) within the Digital Interactions Lab at the University of Amsterdam. I am interested in designing for empathic, human-centred smart buildings that cater to its occupants physical and mental well-being. I have experience in user studies for affective computing as well as independent novel research in neutrino particle detection using deep learning.
I am keen on working at the intersection of Affective AI and HCI. Particularly, I am interested in bringing unobtrusive, ethical and inclusive technology.
MSc Computer Science: Big Data Engineering [Sept 2018 - December 2020]
Joint degree by the University of Amsterdam and Vrije Universiteit allows for study of technology from which data infrastructures are built, allowing for design and operation solutions for processing, analyzing and managing large quantities of data.
BA Arts: Honours in Mathematical Economics with Co-op [Sept 2012 - Apr 2018]
Joint programme offered by the Faculty of Mathematics and the Department of Economics includes the use of differential calculus, differential equations, and mathematical optimization to understand and predict economic behaviour. Studies complement advances in mathematics using problems that come up in economics.
Rao, S. , Alavi, H., & Good, J. (2023, April). Towards Empathic Buildings: Exploring How Smart Buildings May Be Designed to Address Occupants’ Subjective Needs. EmpathiCH: Unravelling Empathy-Centric Design In ACM International Conference Proceedings Series, presented at CHI 2023 Workshop EmpathiCH
Rao, S., Ghosh, S., Pons Rodriguez, G., Röggla, T., El Ali, A., & Cesar, P. (2022, September). Investigating Affective Responses toward In-Video Pedestrian Crossing Actions using Camera and Physiological Sensors. In 14th International Conference on Automotive User Interfaces and Interactive Vehicular Applications (pp. 226-235).[DOI]
Rao, S. , Resendez, V., El Ali, A., & Cesar, P. (2022, July). Ethical Self-Disclosing Voice User Interfaces for Delivery of News. In 4th Conference on Conversational User Interfaces (pp. 1-4). [DOI]
Rao, S., Resendez, V., El Ali, A. & Cesar, P. (2022). Ethical Self-Disclosing Voice User Interfaces for Delivery of News. presented at CHI 2022 Workshop CUI@CHI: Ethics of Conversational User Interfaces. [pdf]
Ghosh, S., Pons Rodriguez, G., Rao, S., El Ali, A., & Cesar, P. (2022). Exploring Emotion Responses toward Pedestrian Crossing Actions for Designing In-vehicle Empathic Interfaces. [DOI]
ICT4D KasaDaka Weather Prediction for Sub-Saharan Africa
Managed an interdisciplinary team towards the development of a voice-based weather service with automated geo-location for farmers in remote West Africa.
Modelled the pr ototype around Human Computer Interaction (HCI) ideals for improved user functionality by using GPS based auto detection of the users location to provide rainfall and storm information with minimal user interactions.
Development was done using Django-Heroku and KasaDaka Voice Service Development Kit (VSDK) within the context of limited technology and diverse cultural and socio-economic aspect in West Africa.
The service is due to be tested in Ghana for further feedback and development as part of the ICT4D research programme at the Vrije Universiteit.
3D Kadaster Urban Landscape of The Netherlands
As part of a 3 person team, developed a 3D model of all the buildings in The Netherlands using 2TB point-cloud data and 18GB 2D building polygons.
SurfSara's Hadoop Cluster and LasPy were used for data pre-processing. Pre-processing techniques achieved significant data reduction, from 2TB to 3MB and 18GB data to 1.5GB respectively.
The 3D point cloud data was represented as octrees and masked and intersected with 2D building layouts.
three.js was used to generate meshes and render the buildings.
Project received recognition in the Large Scale Data Engineering Hall of Fame.
OmniArt Sentiment Analysis and Visualisation
Conducted sentiment analysis using SentiStrength and Python on the OmniArt dataset - a detailed collection of 1 million artworks and their metadata.
Remarks made on various artworks gathered from 17BCE until 2012 were represented as human sentiment values. The goal was to observe the relationship between human sentiments and art across different cultural time periods.
Key findings noted association of the Pop Art era that displayed vivid colours with positive emotions while both the Renaissance and Baroque era were associated with neutral to negative emotions.
Data processing challenges were identified due to quality of the artwork descriptions that were found to be very sparse in nature and often unrelated or inconsistently worded.
CommonCrawl Knowledge Extraction Engine
As part of a team, implemented a knowledge extraction engine to extract entities from the CommonCrawl dataset and linked them to Freebase.
WARC files were streamed with warcio and pre-processing and text mining were performed on the DAS4 cluster with PySpark.
Additionally, a novel method for extracting relations between two entities was proposed for the NLTK Python library. PATTY was used to train a model that predicted the relation between two entities given some filler text with 42.5% accuracy.
User Activity Prediction for Quantified Self
As a part of a team, used the ActiTracker dataset to obtain user activity and predict and classify activity using personal and impersonal models.
Achieved high accuracy of 96% with the personal model with Python.
Co-Chairperson Informatics Institute PhD Council [February 2023 - Present]
Student Volunteer at CHI 2023 [April 2023]
Summer School on Usable Privacy and Security [01 May 2023 - 05 May 2023]
Research Engineer [June 2021 - June 2022]
Working within the Distributed Interactive Systems (DIS) group to assist Prof. Pablo Cesar and Dr. Abdallah el Ali with research in affective computing and AI.
Assisting the group with publishing scientific papers within the domain of HCI and affective computing for workshops and conferences.
Developed project proposals within the scope of Conversational User Interfaces(CUIs) and affec- tive computing and AI.
Research Intern [June 2020 - October 2020]
Investigated application of point-based deep learning for neutrino particle research as part of ongoing Masters thesis in collaboration with the Netherlands e-Science Centre and Nikhef, the Dutch institute for sub-atomic particle physics.
Implemented novel representation of data as 3D meshes for identifying neutrino particles amidst large background noise.
Developed a data processing pipeline and extended the PointNet neural network using the PyTorch framework to classify data containing neutrinos.
Trained with 3 AMD MI50 GPUs on the Nikhef cluster demonstrated network’s ability to identify 95% of the neutrinos. These metrics set the research baseline as the project is the first of its kind.
Demonstrated successful application of non-linear regression techniques to infer energy properties from identified neutrino events with 80% accuracy.
Revenue Analyst [Apr 2015 - Jan 2016]
Developed 13 monthly dynamic business summaries using VBA, PowerPivots and PowerBI tools in Excel for all Microsoft divisions.
Created a 12 month revenue forecast report for Microsoft Office by analyzing 2 GB of data from the Revenue Database using MySQL.
Developed and presented a detailed analysis of inconsistencies in the Azure Cloud billing system to the President of Finance, Canada that was implemented.
Business Analyst [Jan 2014 – Apr 2014, Aug 2014 - Dec 2014 ]
Created 3 bi-weekly reports using Excel and Salesforce to observe performance of Business Development team members.
Measured and presented bi-weekly metrics to measure student employment over the term at the University.
Designed report using Excel to track conversions from job posting to hiring for all University Students.
Created metrics to track and analyse job posting trends across all University Programs. This report was used by the Business Development team to attract employers to hire from programs with lower job postings.
Data Mining Techniques [April 2020 - May 2020]
Supervised the computer lab group of 54 Masters students through completion of a Data Mining project and various assignments.
Conducted weekly meetings to instruct and assist students in their progress through the assignments and addressed machine learning concepts as required.
Languages Python, JavaScript, R, Scheme, SQL, Shell, LaTeX
Statistics Linear Models for Regression, Linear Classification, Perceptron, Linear Discriminant Analysis, Logistic Regression, Neural Networks,Principal Component Analysis, k-Means, Mixtures of Gaussians, Kernel methods, Support Vector Machines, Gaussian Processes
Mathematics Multivariate Calculus, Orthogonal projections, Least-squares, Determinants, Singular-Value Decomposition, Higher Order Linear Ordinary Differential Equations, Black-Scholes Partial Differential Equations, Stochastic Differential Equations, Combinatorics & Optimisation (Solving linear problems, computational complexity, Geometry, Linear programming duality, cutting planes, branch & bound, Non-linear convex optimization)