Professional Experience

Kennesaw State University – Graduate Researcher
Supervisors: Dr. Jiho Noh and Dr. Md Abdullah Al Hafiz Khan
Duration: Jan 2022 – May 2023
Location: Atlanta, Georgia, USA.

Research Project 1: Graph Neural Networks (GNN)

  • I implemented a state-of-art neural fusion network model which transforms streaming data into Graph representations.
  • We implemented GNN and Clustering Techniques on Information Retrieval to optimize the search functions in real-world applications to retrieve the documents according to certain filtering aspects of the user.
  • Migrated PubMed data to Neo4J cloud database to integrate the XML data into parabolic GraphQL structure.

Research Project 2: Open Brain Communication Interface (OpenBCI)

  • Researched Brain Communication Interface (thoughts converted to text) and developed a GUI through PyQt and Tkinter UI design interface to extract Electroencephalogram (EEG) streaming signals as raw data from OpenBCI headset. Further, a state-of-art Attention ML model was developed and integrated into the data generated from GUI. 
  • Experienced and explored the chunking of streaming data through Kafka-Python—hands-on experience on PyKinesis (Amazon Kinesis service) to process the Brain-EEG- Streaming data. 
  • In future work, the brain-to-text data integration can also be made as a continuous integration pipeline using AWS ML platform services to create an interactive development in the real world. 

Research Project 3: Conversation Search and Generative Language Models (GLM)

  • Researched Conversational Information Retrieval (CIR) search systems with supporting concepts like Computational Linguistics, Natural Language Processing, and Dialog Management Systems.
  • Published a research paper at the ACMSE conference on Conversational Search and Applications in Bio-Medicine. Also discussed is GLM, which describes the functioning of interactive GPT-3.5 and ChatGPT. 

Tata Consultancy Services (TCS) – System Engineer
Duration: Dec 2018 – Dec 2021
Location: Hyderabad, India.

Project 1: ASML (Advanced Semiconductor Material Lithography)

ASML is a chip-design company that works on huge data integrations and operates on product life-cycle management. The project aims to provide IT services like Application Portfolio Management, Cyber Security Operations, and Data Management Analysis.

  • Responsible for integrating, migrating, processing, and deploying the data to ASML client servers. As the data is static, the operations on data were taken place through SQL databases and for grooming & planning the sprints Agile Methodologies are implemented.
  • Responsible for ETL operations through Exploratory Data Analytics (EDA) to integrate ASML data. The data was processed through BI tools like Power BI or Tableau for visualizations and analysis to make certain conclusions. 
  • Further, the processed data has to be deployed and integrated continuously into the ASML server, where the utilization of GIT repositories acts.

Project 2: Automatic Video Annotations

Automatic Video Annotation has been proposed to act as a mediator in content-based video Retrieval (CBVR) applications and video management. This study is based on the machine learning technique to annotate videos. Hence, the critical idea of Automatic Video Annotation (AVA) is to construct the model(s) by automatically learning semantic concepts from many videos (even shots or keyframes), then utilize these concept models(s) for predicting appropriate annotation/label for any new video almost like Object Rekognition in AWS cloud.

  • I am experienced in relevant skills, likely Python, Data Structures, Deep Learning, Open CV Packages, Convolutional Neural Networks (CNN), YOLO Models, TensorFlow.
  • Developed Deep Learning Convolutional Neural Networks (CNN) and YOLO Models for Object Recognition. Managing the streaming data without distortion in Image frame processing and achieved this through  Kafka-Python.
  • Analyzing the time-complexities, accuracy, and performance of the model. Though CNN accuracy is low compared to SVM, the time complexity (speed) of CNN is compelling, which is preferable for security—further explored the time complexity rates of different algorithms in ML.

Bharat Sanchar Nigam Limited (BSNL) – Industrial Trainee
Guidance: Sri K. Satyanarayana Raju (Deputy Engineer)
Duration: May 2019 – Jul 2019
Location: India

  • Researched on BSNL cyber-attack that affected south India in Jul 2017. This virus reportedly affected 60,000 modems and blocked internet connectivity. The data was stored in the database for pace evaluation and multiplexing initiations. In the research, the amalgamation of the hypothesis was that the primary leakage would be from their intranet portals. 
  • According to the Bug Bounty program launched by IIT Guwahati, I performed the program on the provided data and created a visual representation of multiplexing communication transformations. Further, I tried various techniques, and the Data Erasure gave me a reasonable accuracy rate. Data Erasure is a method for eliminating a computer file by replacing it with other information. And it allows the hard drive to remain a reusable asset.