R. Prashanth

About Me:

I did my PhD in Electrical Engineering from IIT Delhi. My research topic was on early detection of Parkinson's disease through multimodal data analytics. Prior to this, I did my Masters from IIT Kharagpur in Medical Imaging and Image Analysis.

Presently, I am working as a Lead at Siemens Healthineers (previously Healthcare) where I am involved in different data science works. Prior to this, I was working as a Lead at Fidelity Investments.There as well, I was working on the data science problems for the company. Prior to this, I was working as a Consultant (Data Scientist) for KPMG Global Services Pvt. Ltd. Prior to joining KPMG, I was working as the R & D Manager at Flytxt Mobile Solutions Pvt. Ltd in Thiruvananthapuram, Kerala. Here, I was mostly involved in projects for data analytics which include Social Media Data Mining, Natural Language Processing, Machine Learning, and Big Data Analytics. Prior to this, I was working at Azoi Mobile Technologies as a R and D Engineer. Here, I was responsible to develop predictive models for estimating the blood pressure, using features derived from various signals from the body, through machine learning.

I am passionate in the areas of Healthcare Informatics, Data Analytics, Predictive Modeling, Machine Learning, Deep Learning, Natural Language Processing, Image Processing, Data Mining, Statistical Analysis, and Big Data.

Education:

➢ Ph. D. in Electrical Engineering from the Indian Institute of Technology (IIT) Delhi (July 2010 - Aug 2015)

➢ M.Tech in Medical Imaging and Image Analysis from the Indian Institute of Technology (IIT) Kharagpur (July 2008 - July 2010)

➢ B. Tech in Applied Electronics & Instrumentation Engineering from the College of Engineering Trivandrum, Kerala University (July 2004 - July 2008)

Professional Experience:

• Lead, Siemens Healthineers, Bengaluru, Karnataka, India (Oct 2019 - Present)
• Lead, Fidelity Investments, Bengaluru, Karnataka, India (Jan 2018 - Oct 2019)

At Fidelity, I have been involved in the following projects

1. Call volume estimation

In this work, the aim was to estimate the volume of call on a day. Different techniques such as ARIMA were used for the modelling. The time series data along with other relevant data which is in the form of a hive table is loaded in a Python environment and then analysis is carried out.

2. Customer Intent Modelling

Here, the goal is to estimate the intent of a customer from calls and it is a multilabel classification problem as the same call could have multiple intents. Neural network based models were developed for this task.

• Consultant, KPMG Global Services Pvt. Ltd., Bengaluru, Karnataka, India (July 2017 - Dec 2017)

I am working as a data scientist for the company. At KPMG, I was involved in the delivery of the following projects

1. Anomaly detection on financial transaction data

In this work, data was extracted from SQL database and then prepossessed. After this, algorithms such as isolation forests, k-nearest neighbors and density estimation techniques were used for the modelling.

2. Check of data quality that gets populated in the database

In this work, I had to write SQL queries to check the goodness (quality) and consistency of the data that was getting loaded in the database.

• R & D Manager, Flytxt Mobile Solutions Pvt. Ltd., Trivandrum, Kerala, India (March 2016 - July 2017)

Here I am responsible to carry out data science and machine learning problems for the company. Few of the projects that were successfully carried out are:

1. Topic Extraction from Social Media Data through Distributed Machine Learning

In this work, social media data was analysed to create a model that can extract the topics using the Latent Dirichlet Allocation (LDA) technique. Spark programming was used to create a distributed model.

2. Sentiment Analysis of Social Media Data

Here a model for analysing the sentiment of the social media data was developed. Opinion-lexicon based approach was used here.

3. Most Popular / Top Trending Item Computation from Social Media Data

In this task, a model to estimate the top trending item was computed by finding the number of occurrences.

4. Trend Analytics for Predicting Medals

Here a model to predict the medals for a country using trend analytics techniques of simple moving average and weighted moving average is carried out.

5. High Accuracy Predictive Modeling of Customer Churn

In this work, predicitve models for predicting customer churn is carried out.

6. Deep Learning for Superior Customer Relationship Management

Deep Learning finds immense applications in different areas of machine learning. In this work, deep learning models including deep neural networks, deep belief networks and recurrent neural networks were used.

7. Social Media Analysis to Generate Insights for Kerala Elections

In this work, social media data from relevant pages were extracted and analysis using k-means was carried out.

• R & D Engineer, Azoi Mobile Technologies Pvt. Ltd., Ahmedabad, Gujarat, India (Jun 2015 - Feb 2016)

I was responsible to develop predictive models for estimating the blood pressure from biosignals that were acquired from a subject using a mobile device. The biosignals, namely Electrocardiogram (ECG) and Photoplethysmogram (PPG), signals extracted were processed via signal processing techniques such as peak detection, to extract features, followed by using machine learning methods to create predictive models. Various techniques such as linear regression, ridge regression, neural networks and random forests were tried.

Research Experience:

• PhD Thesis, IIT Delhi (July 2010 - Oct 2014)

Dissertation Title: Computer-Aided Early Detection of Parkinson’s Disease (PD) through Multimodal Data Analysis

Early detection of PD is a challenging and an important problem. It is crucial for early management and for better treatment regimens. In this work, various analytics and predictive modeling using multimodal data has been carried out for the early detection of PD. The data used for the study are neuroimaging data (SPECT and MRI), clinical examination data and biological (cerebrospinal fluid) data. Image processing was carried out to process the images and extract the features, and along with clinical and biospecimen data were used to develop predictive models. Machine learning and statistical tests were used to analyse the features. Various techniques such as logistic regression, Naive Bayes, support vector machines, random forest were used in the analysis. The work has been published in many international journals and conferences. One of the published works was among the winners of the Best Paper at the IEEE EMBC 2014 conference (A prestigious conference of the IEEE).

• M.Tech Thesis, IIT Kharagpur (July 2009 - May 2010)

Thesis Title: Characterization of Choroidal Neovascularization (CNV) through Analysis of Fluorescein Angiograms and Optical Coherence Tomograms

Advisors: Dr. Jyotirmoy Chatterjee and Prof. Pranab Kumar Dutta (IIT Kharagpur)

CNV is a disorder that affects the retina of the eye causing vision loss. In this project, characterization of CNV lesion through analysis of retinal images was carried out. Color fundus, Fluroscein Angiogram and Optical Coherence Tomography images were used for the analysis. Quantification of the lesions through image processing, segmentation, region of interest analysis were performed. Novel methods for better characterization of the disease were obtained.

• M.Tech mini-project, IIT Kharagpur (Dec 2008 - May 2009)

Project title: Robust Detection of Microaneurysms in Sight Threatening Retinopathy

The project carries out image analysis for the detection of microaneurysms from fluorescent angiogram images through preprocessing and segmentation.

Teaching Experience:

• Department of Electrical Engineering, IIT Delhi

Computer Vision, Teaching Assistant (Jul 2011 - Dec 2011, Jul 2012 - Dec 2012)

I was responsible to evaluate the course assignments and review program codes of student groups.

Honors and Achievements:

• Among the Best Paper Award Winners in BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Research at IEEE EMBC’14.
• Individual award at Fidelity for developing a machine learning model for customer intent modelling in Python.

As a Reviewer:

• Neurocomputing (An Elsevier journal)
• Neuroscience (An Elsevier journal) - An outstanding reviewer for the journal.
• Engineering Applications of Artificial Intelligence (An Elsevier journal)
• IBRO Reports (An Elsevier journal)
• Molecular Imaging and Biology (A Springer journal)
• International Journal of Imaging Systems and Technology (A Wiley journal)

Resume/Curriculum Vitae:

My resume is available here.

Social Presence:

• LinkedIn
• Quora: You can see my answers on Quora here.

Some of my most viewed answers:

1. How do I write a research proposal?
2. What is the exact procedure for publishing a paper in IEEE?
3. How can I find my PhD dissertation topic?
4. If I have a research paper how and where can I publish it?
5. What is supervised learning?
6. What is the difference between machine learning model and ML algorithm?
7. What are the differences between bagged trees and random forests?
8. I am using SVM to train a dataset, but when I test the model on the same dataset it does not give 100% accuracy. What is the problem with my model?
9. What is the main difference between a SVM and SVR?
10. What is one hot encoding and when is it used in data science?

Contact Information:

• Feel free to send an email to prashanth.r.iitd@gmail.com