My research interests primarily lie in Natural Language Processing (NLP) and Large Language Models (LLMs). I examine the complexity of Long Distance Dependencies (LDDs) observed in sequential datasets and their impact on the representational capacity of Large Language Models (LLMs) in an attempt to make them more efficient. I also work in the areas of Deep Learning, Linguistics, Computer Vision, and Information Theory. I am conducting my research under the guidance of Prof. John D. Kelleher at Technological University Dublin, Ireland.
Previously, I worked with National Institute of Oceanography, Goa, India, where I developed and upgraded Autonomous Underwater Vehicles (AUVs). I co-founded an IoT company, SpitiQ, Goa, India, where I was instrumental in building a home automation solution using Wireless Sensor Nodes (WSN). I also assisted Cyanodoc Healthcare, Goa, India, in developing a doctor’s appointment application.
PhD in Deep Learning (NLP), 2023
Technological University Dublin, Ireland
BE in Electronics and Telecommunications, 2010
Goa Engineering College, Goa, India
Designed the web application architecture of the doctor’s appointment system and deployed it using Amazon AWS.
Advised on the development of the Android and iOS apps of patient’s and doctor’s user interface.
Advised on the design of the differential diagnosis system using Bayes’ theorem for endocrine diseases.
Consulted with medical experts to improve the differential diagnosis system by optimizing the probabilities of the symptoms manifesting due to the presence of a disease.
Designed and manufactured wireless sensor nodes or motes and a border router using Atmel’s Atmega128RFA1 microcontroller that includes a wireless transceiver onboard compliant with IEEE 802.15.4 stack and 6LowPAN protocol with mesh connectivity for home automation. Ported Contiki OS on the mote.
Integrated smart home sensors and electric switching capability with the motes.
Designed a communication protocol on MQTT running on Contiki OS on the mote to enable communication between motes and the broker on AWS IoT Core. Designed an Android application to view the sensor and switch data and send instructions to the motes.
Developed and maintained two robots (i) Autonomous Underwater Vehicle (AUV-MAYA) and (ii) Autonomous Vertical Profiler (AVP), at Marine Instrumentation Division. Participated in field trials of AUV-MAYA and AVP.
Improved the wireless communication of AUV-MAYA by modifying communication protocol on TCP/IP.
Developed a Hardware-In-Loop (HIL) Simulator for AUV-MAYA.
Developed a satellite communication portal for communication between AVPs deployed in remote locations (with no Wi-Fi or mobile network) and the servers located in the lab, to transmit the status and data of the AVPs and display their status on the website.
Developed graphical user interfaces on Windows and Android for AUV-MAYA and AVP to perform their missions.
Developed a mission planner for AUV-MAYA that allowed the operator to plan mission paths for AUV-MAYA on Google Earth and converted them into mission files to be uploaded on AUV-MAYA.
Developed websites with content management system (CMS) for Open Source Drug Discovery, Bangalore, India, and Bharatiya Vidya Bhavan, Goa, India.
Developed a program for the Raspberry Pi to enable the developer from developing applications on Raspberry Pi without any hardware or programming knowledge.
Conducted tutorials on CUDA programming at Goa Engineering College, Goa