profile-pic

Shubhashri Rane

As an applied research scientist at Intel, I work on developing and deploying cutting-edge AI and NLP solutions to enhance the performance and functionality of Intel's products and services. I have 7+ years of industry experience in software engineering and machine learning, during which I have worked on a blend of development and validation automation tasks.

I have a strong academic background in machine learning and AI, with a master's degree from Liverpool John Moores University and a postgraduate diploma from the International Institute of Information Technology Bangalore. I am proficient in Python, Java, and Perl, and I have hands-on experience in using TensorFlow, Pytorch, NLTK, Scikit-Learn, and other data science libraries and frameworks. I am passionate about learning new technologies and applying them to real-world problems. My goal is to contribute to the advancement of AI research and innovation.

  • Role

    Applied AI Research Scientist

  • Years of Experience

    6 years

Skillsets

  • Bash scripting
  • Visual Studio
  • Unix/Linux
  • Tableau
  • Scrum
  • Scikitlearn
  • Git
  • Anaconda
  • Agile methods
  • Jenkins - 2 Years
  • Java
  • TensorFlow
  • SQL - 3 Years
  • PyTorch
  • Python - 3 Years
  • Perl
  • pandas
  • NumPy
  • Matplotlib
  • LangChain

Professional Summary

6Years
  • Apr, 2020 - Present5 yr 11 months

    Applied AI Research Scientist

    INTEL
  • Feb, 2019 - Mar, 20201 yr 1 month

    Software Engineer

    INTEL

Applications & Tools Known

  • icon-tool

    Anaconda

  • icon-tool

    Visual Studio

  • icon-tool

    Git

  • icon-tool

    Jenkins

  • icon-tool

    Tableau

  • icon-tool

    MS Excel

  • icon-tool

    Scrum

  • icon-tool

    Tableau

  • icon-tool

    MS Excel

Work History

6Years

Applied AI Research Scientist

INTEL
Apr, 2020 - Present5 yr 11 months
    Led end-to-end development of a code generation tool; developed preprocessing techniques for RTL code; designed ML solutions for regression prediction and test deduplication; developed chatbot with Haystack.

Software Engineer

INTEL
Feb, 2019 - Mar, 20201 yr 1 month
    Contributed to plugin and functional validation; accelerated bug fixes; implemented performance testing; restructured internal product documentation.

Achievements

  • Led the end-to-end development and optimization of a code generation and explanation tool, involving prompt engineering and meta prompting for RTL code.
  • Transformed the codebase using LangChain to adopt a sequential chain architecture, integrating streaming capabilities to enhance user experience, resulting in a 30% increase in processing efficiency and significantly streamlined workflows.
  • Developed preprocessing techniques for RTL code, ensuring high-quality inputs that enhanced the accuracy and efficiency of our machine learning tools.
  • Spearheaded the development of an ML solution to predict regression run time for post-silicon tests to identify bad runs at an early stage which in turn will propagate savings of $9.9K per quarter.
  • Built a model for test case de-duplication problem using Sentence Bert and DBSCAN clustering technique for identifying similar post-validation tests which lead to a reduction in validation cycle by >20% per platform program in turn reducing the execution costs.
  • Developed a chatbot using Haystack, enhancing customer interaction and service capabilities.
  • Accelerated over 40% of bug fixes by reproducing, testing, and documenting production bugs.
  • Successfully implemented Performance Testing during production releases of the application which gave >90% insight of production scaling.
  • Spearheaded the restructure of internal product and process documentation which enabled faster onboarding for new hires leading to accelerated solutions for customer queries.
  • Directed the interaction and partnership between business units to identify and define key business problems, ensuring active cooperation and alignment with project goals.
  • Built a model for test case de-duplication problem using Sentence Bert and DBSCAN clustering technique for identifying similar post-validation tests which led to a reduction in validation cycle by >20% per platform program in turn reducing the execution costs.
  • Core contributor in the plugin and functional validation for Intels in-house storage management tool which includes automating and validating over 150 test-cases for high impact features and bugs using Java.

Major Projects

5Projects

Brain Tumor Segmentation using Deep Learning

    Evaluated architectures for brain tumor segmentation from MRI scans; analyzed transformer-based techniques compared to Unet architecture.

Sentiment Based Product Recommendation System

    Built sentiment-based product recommendation system; developed data sourcing, sentiment analysis, and deployed UI using Flask on Heroku.

Telecom Churn Model

    Performed EDA and feature engineering to identify key attributes affecting customer churn in the Telecom industry.

Gesture Recognition

    Recognized gestures from video using 3D convolutional networks for Smart TV.

Navigation Aid for Visually Impaired People

    Engineered prototypes to aid visually impaired people using Raspberry Pi and image processing.

Education

  • Master of Science in Machine Learning and AI

    Liverpool John Moores University (2022)

  • P.G. Diploma in Machine Learning and AI

    IIITB (2021)

  • B.E. in Computer Science and Engineering

    SDM College of Engineering and Technology (2018)

Certifications

  • Machine learning specialization from stanford university