Sruthikeerthi Rajagopalakrishnan Email & Phone Number

New Jersey, United States

United States

Omdena FINZ Project: This project involved developing a pipeline to include geopolitical and climatic factors while scoring a proposal as feasible or non-feasible.A pipeline which took numeric inputs to identify geopolitical risks and image data to identify climatic risks which then gave a score to the proposal based on the location input was designed.I was an active collaborator in the data collection task where we used image sources such as Sentinal 1, 2 and 2A to collect satellite images for modelling , and the image modelling task were we developed image segmentation models to identify flood prone areas.Omdena GlintSolar Project : This project attempts to develop a solution to the scaricity of land to install solar panels to reduce global warming. The pipeline developed essentially eliminates noise in the the satellite image during the depth estimation process to determine if a piece of land is feasible to install solar panels.I was an active collaborator in the data collection task where we collected satellite image data to identify the water depth and other affecting factors precisely.Omdena Hashiona Project: This project attempts to find correlations, symptoms and morbidity for Hashimoto’s Disease. I have been selected as a part of this team and this project will commence in February 2022.

Omdena WRI Land Management Project: This project focused on developing a Knowledge Graph for the unstructured data from gazatte and court documents to solve land misappropriations in Kenya. The raw data was collected, annotated, processed to find the important entities and then a Knowledge was constructed using the recognized entities and relationships.I was the task leader for the Named Entity Recognition task and an active participant in the Annotations task. Along with my team, I developed models to detect the major named entities such as places, numbers and owner details using Spacy, BERT and Attention based models. We used a custom set of tags specific to our problem statement such as Owner Name, Deed Number, Property ID etc.We observed that the Spacy model gave the best accuracy of 96% and we extracted the entities from about 36000 gazatte documents from the time period 2005 to 2018.

Chennai, Tamil Nadu, India

I worked in the Robert Bosch Center for Data Science and Artificial Intelligence on the project titled Data Analytics based Performance Evaluation, Location, Dispatch, and Routing of Ambulances. My role included designing a dashboard using Dash and Folium to visualisation the performance metrics and coverage of the EMS system, analysing geospatial data to optimise the position and routing of ambulances, and designing optimisation algorithms for efficient positioning and coverage of ambulances across five districts of Tamil Nadu.As a part of this project, an algorithm to detect the incident location of a first response call using the geospatial data was developed. This work won the Best Paper Award to the Conference on Deployable AI 2021 organised by IIT Madras. An extension to this work was presented at ACM CODS-COMAD 2022 and published in the conference proceedings.Further, I worked with the volunteer team on advising the Tamil Nadu COVID 19 War Room team for improving the interventions introduced by the government to stop the spread of the pandemic by developing epidemiology models and visualisation tools.

Bengaluru, Karnataka, India

I worked with the Video Analytics team. I worked on optimizing object detection models using YOLOv4 algorithm, by reducing the precision of the model from Floating Point 32 to Integer 8 using the Post-training Optimization Toolkit in Intel® Distribution of OpenVINO™ Toolkit. I also developed inference engines and performed a detailed benchmarking analysis to compare the inference time for the open-source Tensorflow model, OpenVINO™ FP32 model and OpenVINO™ INT 8 model.

Bangalore

I worked with the Navigation Systems Area at ISRO Telemetry Tracking and Command Network facility. The project focused on developing a stable and robust timescale signal for satellite navigation and communication using an ensemble of atomic clocks. Artificial neural networks were used to model the ensemble. We published a research paper on the same in MAPAN Journal of Metrology Society of India.

Chennai Area, India

A one-month summer internship where I explored Machine Learning and Deep Learning algorithms, techniques in data visualization. I designed recommendation systems using collaborative filtering and unsupervised algorithms using the MovieLens dataset.

Chennai Area, India

A summer internship during which, I explored the intuitive concepts of Cloud Computing and Amazon Web Services(AWS) applications were explored. Along with a team of four, I also developed a serverless website which displays the price and other details of commodities from a sample dataset using AWS services such as S3, DynamoDB, Lambda Compute and API Gateway.

Masters Of Science In Data Science, Engineering, 3.6 / 4

Bachelor Of Technology, Cse, 9.05 / 10

First-class with distinction Relevant coursework: • Core subjects: Data structures and algorithms, object-oriented programming.

Aissce, 93.6

Aisse, 9.8