Operational Data Store

Infosys is a consulting and technology firm with more than 187,000 employees and revenues of US$9.02 bn in FY16 enabling clients from around the world to stay ahead of the innovation curve. Infosys developed Finacle, an universal banking solution to provide banks with a tool to excel in their business. Finacle is used by banks across 84 countries and serves over 450 million customers which is nearly 18% of the world's adult population [1]. Finacle includes 24 products in its portfolio which help clients to improve banking experience for their customers. While at Infosys I primarily worked on Finacle Operational Data Store (ODS). Finacle systems interact with humongous amounts of data which is being generated and stored at the banks with every new transaction. Finacle Operational Data Store is a big data solution to enable banks to unlock the potential of this data to meet the bank’s demand for reporting, risk management, analytics etc. The production databases support millions of transactions each day. These databases are required to have higher latency to support near real-time transactions. The load on production databases is relieved by ODS by taking over the reporting and analytics load. Thus improving the efficiency of the production environment. I worked on the development of the ODS project while it was still in infancy. I worked on defining the mapping between attributes in production databases and ODS so as to confirm with different reporting standards in different countries. This task required me to study the bank reporting requirements and banking processes, terminologies etc in different countries of implementation like UK, South Africa, India etc. Later on I worked on the High level and Low level design for the implementation of various reports like the MIS reporting systems, asset liability management, risk management solutions and relationship-based pricing reports for implementation within ODS. Additionally I developed tools for automating my tasks which were later adopted in the organisation to improve the productivity of teams working on the project. The development of the database schema to build the ODS on top of production databases resulted in large untapped revenue stream for the organisation. Find more details about the project in this brochure

Data Analysis

After completing my postgraduate diploma from the University of Edinburgh I briefly worked with Dr Satyen Deshpande to analyse the clinical data he had collected for his PhD thesis. I used Anova for testing the effects of his proposed treatment on patients. The variance analysis lend support to the claim of effectiveness of his proposed treatment methodology. More description…..

Finacle Technical Consultation

Infosys is a consulting and technology firm with more than 187,000 employees and revenues of US$9.02 bn in FY16 enabling clients from around the world to stay ahead of the innovation curve. Infosys developed Finacle, an universal banking solution to provide banks with a tool to excel in their business. Finacle is used by banks across 84 countries and serves over 450 million customers which is nearly 18% of the world’s adult population [1]. Finacle includes 24 products in its portfolio which help clients to improve banking experience for their customers. Finacle is enabling Clarien Bank (formerly Capital G) with its e-banking solution to provide its customers with a robust and fast service. Sigma Global Solutions, CA supported the operations of the Finacle systems deployed with Clarien Bank. I worked with Sigma Global Solutions as technical consultant for the Finacle E-Banking product. My role involved supporting the daily operations of the system, assisting the bank with analysing the new customization requirements and coordinating with the vendor, Infosys to include the new features in their product. I did the consultation two days a week.

Cellular Structure Classification

Oxyent Medical (now AIIT-SPARC) developed an innovative solution that would transform histopathology domain through image analytics. The product is aptly named Image Analysis and Decision Support System (iADSS). iADSS is an HPC based Machine Learning and Image Processing Solution for Histopathology Image Analysis coupled with advanced viewing capabilities. iADSS in its first version aims at helping histo-pathologists by reducing their dependency on cumbersome microscopic analysis and there by expediting the whole process by many folds. This is a niche product which will help us deliver our clients with significant reduction in the drug development cycle time, improved accuracy and efficiency. The drug development cycle in a pharmaceutical company, which start with a concept drug and ending with the formulation becoming a market reality is lengthy; the duration of concept-to-market cycle is usually 8-12 years. Preclinical toxicology studies form an important, but time consuming phase in the drug development cycle. These studies are uniquely identified by Histopathological analysis of organ tissues of animals to determine whether the given tissue was toxically affected by the candidate drug. Computer Aided Decision Support (CAD) system can play a major role in expediting this analysis process and thereby aiding speedier drug availability in the market. The input to our CAD system is multi-resolution images obtained by whole slide scanning of H&E stained animal organ tissue slides. The organ tissue images need to be analyzed using histological, textural and color features at various resolutions to determine if the effect of the drug administered to the animal, to which the organ tissues belong, is normal or toxic. Unlike the CAD systems used in medical imaging scenario where the abnormality in question is predefined, in the pre-clinical context the abnormality is not known apriori. The tissue slides can be scanned using a batch process scanner, to form a multi-resolution image. The multi-resolution Histopathological image, which is on average 50000x50000 pixels, has the humongous data in each of them. This calls for Machine Learning and Image Processing Approaches for highly contextual to pre-clinical toxicology domain, to be implemented in HPC infrastructure. We used image processing to segment the images to identify distinct micro-cellular structures in the images. I worked on the development of the machine learning algorithms using neural networks for the classifying the image segments into respective micro-cellular structures. The work of the team led to the filing of two patents. It was an enriching experience to work on the application of the machine learning algorithms in the field of pharmacology.

Named Entity Recognition

Germin8 is a Social Media Intelligence company focused on helping organisations understand and act in real time on the gazillions of conversations by their stakeholders. Germin8 has a range of products that provide brand monitoring, campaign analysis, customer management, industry analysis etc. Germin8 is leading innovation in social media analytics and has a clientele of many reputed companies. At the heart of every product is an algorithm engine that crawls numerous sources on the internet for relevant conversations and subsequently the natural language processing algorithms are applied for processing the big data obtained from the internet. Machine learning algorithms are then used to identify patterns and generate meaningful insights for the organizations. These insights enable organizations to take appropriate actions to retain advantage in highly competitive marketplaces. I worked with Germin8 on the application of the named entity recognition algorithm in unstructured text, eg., from twitter, facebook, etc. The identified entities can then be attributed with the topic of the conversation, sentiment etc for automatic insight generation which is shared with the clients. Identifying named entities in unstructured text like tweets is far difficult than in structured text like blogs or news. Other than incorrect grammatical constructions tweets and facebook comments use many abbreviations and slang words which present difficulties for automating text processing. To overcome these difficulties was a great learning experience for me and it helped me a lot in later projects.

Future Mobility Sensing

The Singapore-MIT Alliance for Research and Technology (SMART) is a major research enterprise established by the Massachusetts Institute of Technology (MIT) in partnership with the National Research Foundation of Singapore (NRF) in 2007. SMART is the first entity in the Campus for Research Excellence and Technological Enterprise (CREATE) developed by NRF.At SMART-Future Mobility group(FM) we developed the Future Mobility Sensing (FMS) system, a smartphone and web prompted-recall based travel and activity survey system, which uses a smartphone application (available for both Android and iPhone) to automatically log and upload sensor data from user’s phones. The uploaded raw data is processed in the backend to construct the user’s activity diary, i.e., we infer user’s stops, modes of transportation and activities. The activity diary is then presented in a user-friendly web-interface in a prompted-recall manner for user verification. The FMS is a next-generation travel behaviour survey system that leverages pervasive smartphones, advanced sensing and communication technologies, and machine learning architecture. It delivers previously unobtainable range of data reflecting what people do, not what they say they do. We field tested FMS in Singapore in conjunction with Singapore Land Transport Authority’s (LTA’s) Household Interview Travel Survey (HITS) 2012. The test recruited more than 1500 users and produced a large set of rich and detailed travel/activity data that has been verified by the respondents. Since then FMS has been deployed in many different countries to conduct various studies like Happy Mobility Sensing, Mobile Market Monitor , Victorian FMS

Happy Mobility Sensing

Understanding and incorporating measures of travel well-being in transportation research is critical for the design and evaluation of policies aiming at enhancing well-being. In recent years several efforts have been made to quantify travelers’ subjective well-being using travelers’ self-reported state of happiness while participating in various activities or travel patterns. Happy Mobility Sensing is a smartphone-based travel survey to measure activity and travel happiness.We collect data about activity locations using smartphones enabled with GPS, GSM, Wi-Fi and accelerometer; and with this raw data we generate the activity diaries of the individual (which include the trip origin, destination, start and end time, and mode). This data is then made available to the individual through a web interface where the individual can verify his/her trips and activities information, and also answer other questions about his/her satisfaction with particular activities. I worked on the development and improvement of the algorithms for inferring activity and travel information from raw sensor data. The algorithms developed in this project have led to filing of a patent. The work on the algorithms and analysis of happiness data collected in the survey resulted in two conference papers.Know more about the project

Victorian Future Mobility Sensing

This project is part of an exciting new Urban Mobility and Intelligent Transportation initiative by the University of Melbourne, in partnership with Department of Economic Development, Jobs, Transport and Resources (DEDJTR), Massachusetts Institute of Technology (MIT), and Singapore-MIT Alliance for Research and Technology (SMART). FMS Victoria will collect personal travel data using a smart-phone app developed by our partners at SMART. The initiative will provide insights into Victorians’ travel demand, mobility patterns and satisfaction of transport choices. The project will allow urban and public transport planners to consider possibilities for future improvements to the city, including:

• introducing new mobility modes such as demand-responsive transport
• development of self-driving vehicles
• understanding how people move around their local areas.

Busotted

Infocomm Development Authority of Singapore (IDA) organized a hackathon called Data In The City in Singapore. I participated in the hackathon along with Steve Robinson, Rudi Ball, and Balajee Amalvayal Rangarajan (my colleagues at SMART). We developed a prototype travel planner application to encourage use of public transport in Singapore. The prototype was judged winner of the competition.