Experience

• 

  Indian institute of information technology design & manufacturing kancheepuram

  May 2019 - Jul 2019

  Research intern

   Conducted background research, literature survey, gathered various data sources, Implemented various machine learning techniques including ELM, DNN, CNN, Random Forest with AdaBOOST. Proposed three CNN architectures for multi tumor classification of brain MRI.  Designed and Fine Tuned a CNN model to increase the optimum utilization of memory resources and evaluated the model on various metrics and visualized the model metrics and documented the findings. Published a research paper on the tumor classification model in International journal of imaging systems and technology. (https://onlinelibrary.wiley.com/doi/abs/10.1002/ima.22554) Show less

• 

  Centre for development of advanced computing (c-dac)

  May 2020 - Aug 2020

  Intern

  The project work is based on the development of a Deep Learning Model to analyze a given chestof COVID infected person. Developed a prototype of a customized system based on deep CNNsfor the identification of COVID-19 as a classification task. An improvement on VOS and VOT along with CNN-based semi-supervised learning techniques are implemented. Training model is implemented in High Performance Computing (HPC) with GPUs programming environment for NonCOVID and COVID based disease open source data sets. Show less

• 

  Ltimindtree

  Jun 2021 - Jul 2023

  Data engineer

  Worked on a client based POC that involves automating the task of data extraction from paramount invoices using the DOC AI google machine learning tool. The extracted data is processed and exported to structured xml files. Trained an OCR model in DOC AI with an F1 core of 0.99. Then developed a python script that would initiate a dataflow job which converted the json files to xml and stored the results in gcp bucket. My last role involved in a project in which I was responsible for developing apache airflow pipelines that ran in Google cloud composer. The objectives of these pipelines varied depending on the requirement but broadly include data cleaning, data exploration and data processing. I would also analyze and optimize existing pipelines that enabled automation and failure recovery without human intervention. Show less