CV

Education

• Ph.D in Computer Science, New Jersey Institute of Technology, New Jersey, USA. [2023 - Present]

• M.E. in Computer Science & Engineering, Jadavpur University, West Bengal, India. [2017 - 2019]

  Thesis: Volumetric Brain MR Image Segmentation using Entropy based Fuzzy clustering algorithm

  Advisor: Prof. Jamuna Kanta Sing

• B.E. in Computer Science & Engineering, Visvesvaraya Technological University, Karnataka, India. [2012 - 2016]

Skills

• Programming Languages: C, C++, Java, Python, R
• Databases: MySQL, PostgreSQL, Neo4j, Redis, Hive
• Libraries/Frameworks: Flask, FastAPI, Tensorflow, Pytorch, Solr, Elasticsearch, Spring
• Tools: Git, Docker, Kubernetes, Talend, Jenkins, Jira

Projects

Vector Databases for Data Mesh

Integrated MLOps Platform

Data Federated Services: Customer 360 Platform

Contextualized Knowledge Graphs

Market Share Prediction

Human Resource Data Analysis

Logistics Demand Analysis

COVID-19 Q&A System

Data Ingestion and Topic Modelling

Gaussian Energy-based function for detection of moving objects in videos

Entropy based Fuzzy clustering algorithm for MR image segmentation

Deep Learning through Image Analysis of real-time videos

Automated Driver Assistance Systems

Academic experience

• New Jersey Institute of Technology:
  • Teaching Assistant - CS 301: Introduction to Data Science (Instructor: Prof. Kaustubh Vijaykumar Pethkar) [Spring 2024]
  • Teaching Assistant - CS 104: Computer Programming & Graphics Problems (Instructor: Prof. Mohit Dale) [Fall 2023]
• Jadavpur University:
  • Teaching Assistant - CSE/S/311: Computer Graphics Lab (Instructor: Prof. JK Sing & Prof. Subhadip Basu) [Aug - Dec, 2018]

Industry experience

• Accenture:

  • ML Engineering Associate Manager:
    • Developed a custom library for interacting with vector databases and knowledge graphs that can interface with a Data Mesh.
    • Improved language model training time and pipeline deployment (by almost 20%) by designing custom ML code templates using LangChain and vector databases.
    • Designed an integrated a zero code MLOps platform by creating custom code and model boilerplates of various machine learning algorithms in TensorFlow and Spark Mlib having seamless integration with their existing infrastructure.
    • Improved model training and deployment times by 31% and 48% respectively from the previous manual deployment process in the K8s workspace.
  • ML Engineering Lead:
    • Developed a data federated platform with the scope of providing integrated customer recommendation experience. BERT fine-tuned on a custom dataset is used for recommendation (validation accuracy ~91%). Production platform handles data at a scale of over 10TB daily with the scope of increase in the future.
    • Designed a Data Quality Engine for indentifying data issues at the source level and using different machine learning algorithms to suggest possible remedies. Multiple ML models and NLP approaches are being used in this project.

• BRIDGEi2i Analytics Solutions:

  (BRIDGEi2i Analytics Solutions became a part of Accenture from 1st May 2022)

  • Lead ML Engineer:
    • Developed data federated platform with the scope of providing integrated customer recommendation experience. BERT fine-tuned on a custom dataset is used for recommendation (validation accuracy ~91%). Production platform handles data at a scale of over 10TB daily with the scope of increase in the future.
    • Implemented a labeled property graph for representing contextual information across various documents. Fuzzy searching algorithms with graph neural networks is used for achieving better properties on node values.
  • Data Engineer II:
    • Analyzed time series logistics data for detecting anomalies across 15 (key performance indicators) KPIs, along with determining their causal relationships and a bi-weekly data refresh of around 2GB.
    • Set up ingestion pipelines for handling monthly data refresh of roughly 25GB and detected change points for identifying a level shift in the time series of HR data.
    • Detected anomalies and forecasted trends within a given window frame on market sales data having 28 KPIs and monthly data refresh of approximately 3GB.
    • Designed and developed the question answering module along with summarization using BERT for a COVID-19 response dashboard on top of the COVID-19 Open Research Dataset.
  • Business Analyst:
    • Worked on building data ingestion pipelines and performed Topic Modeling along with Sentiment Analysis for categorizing text data.
    • Designed a proof-of-concept(POC) for tracking people’s movement on live streaming videos.

Publications

A New Entropy Based Fuzzy Clustering Algorithm for Volumetric Noisy Brain MR Image Segmentation

Certifications & Awards

• Certifications:

  • Natural Language Processing Specialization by DeepLearning.AI from Coursera [20 Feb, 2021]
  • Deep Learning Specialization by DeepLearning.AI from Coursera [02 Aug, 2020]
  • Preparing for Google Cloud Certification: Cloud Data Engineer from Coursera [29 Apr, 2020]

• Selected awards from Accenture:

  • Innovation Award for designing Vector Databases implementations [Jun. 2023]
  • Client Collaboration & Contribution Award for successful growth stories [Dec. 2022]
  • Lead Award for guiding the team through technical and functional challenges [Aug. 2022]
  • Evangelist Award for innovation in annual innovation forum [Dec. 2021]
  • Individual award for above and beyond performance in the year [Mar. 2021]
  • Multiple Team Awards (awarded to the whole team) [Apr. 2022, Aug. 2020, Sep. 2019]