About me

I am Pooja Parab, a Data Science Professional driven by a passion for solving intricate problems using data. Currently pursuing my Masters in data science at Indiana University Bloomington, I derive immense joy from implementing machine learning algorithms, grasping the core data and statistical nuances that power them, comprehending deep learning neural networks, and constructing robust data pipelines.

My journey spans roles as a Machine Learning Intern, Data Science Research Assistant, and Senior Data Engineer across organizations like Quantiphi, Thrivent, and the O'Neill School of Public and Environmental Affairs. With experience traversing internships, research ventures, and industry responsibilities, I excel in transforming projects from ideas into tangible realities.

I firmly believe in the potency of "The art of storytelling is reaching its climax with the emergence of data-driven narratives." Contributing to the narrative through data-infused solutions fills me with satisfaction. As I continue on this path, I aspire to etch my mark by weaving data and narratives into a compelling tapestry that leaves a lasting impact.

What I'm Good At...

  • Model development icon

    Model Development

    Creating intelligence from data through iterative learning.

  • MLOps

    MLOps

    Streamlining machine learning workflows for efficient deployment.

  • data engineering icon

    Data Engineering

    Architecting robust data pipelines for effective data utilization.

  • Statistics

    Statistics

    Embracing the beauty of patterns through numbers.

testimonials

  • Pratik Bijam

    Pratik Bijam

    Pooja is fantastic person to work with & thorough in everything she does. When Pooja is working on a deliverable one can observe her qualities like dedication, motivation, ownership, etc. She is a quick learner and always look out for new learning opportunities. There has never been a time when she has left our team without a solution. Pooja not only has wide range of technical skills but she is also a really good mentor to the junior team members. Her journey from a fresh graduate to lead the team is very exceptional, I would not hesitate a single second to recommend Pooja for challenging assignments.

  • Pushpalatha Sekar

    Fascel Fernandes

    Pooja demonstrated good understanding of ML Engineering workflows. The contribution that she made for model monitoring implementation for few of our models has been great. She also works with the a positive and open mindset to learn new technology, take up new challenges to explore. She is a great team player and was able to collaborate with many team members within in a short span of time. If the right opportunity comes in, i would like her to be part of our team once again

  • Fascel Fernandes

    Fascel Fernandes

    Pooja is one of the most sincere and dedicated team member I have worked with it. I had a chance to work together with her in Quantiphi as track leads and I can definitely say that she is an awesome team player and team lead who has a great ability to solve complex problems and encourages her team to do the same. I would always recommend Pooja if you are looking for a highly skilled and efficient team member!

Worked at

Clients

Resume

Education

  1. Indiana University Bloomington

    Aug 2022 — May 2024

    Master of Science in Data Science
    CGPA: 3.8/4
    Relevant Coursework: Applied Machine Learning, Deep Learning Systems, Advanced Natural Language Processing, Algorithms, Data Mining, Advance Database Concepts, Statistics, Applied Database Technologies.

  2. Dwarkadas J. Sanghvi College Of Engineering

    Jul 2015 — Jun 2019

    Bachelor of Engineering (Electronics Engineering)
    CGPA: 3.7/4
    Relevant Coursework: Object-Oriented Programming, Structured Programming, Computer Networks, Digital Image Processing

Experience

  1. Machine Learning Intern, Thrivent Financials

    Jun 2023 — Aug 2023

    Skills Showcased: Databrick, AWS, Bamboo, Bitbucket, MLOps, Machine Learning, Airflow
    I played a key role in creating a monitoring framework for mutual fund propensity models. This framework covered base statistics, concept drift detection, and quality checks. This effort led to substantial reductions in potential financial losses by preventing faulty model training and quickly alerting stakeholders about misclassifications and performance problems.
    I took charge of putting the same framework into action for production. Using Bamboo, Airflow, and Bitbucket, I ensured a smooth deployment process that seamlessly integrated with our existing system.
    I also conducted research that resulted in a Proof of Concept (PoC) for implementing a Large Language Model (LLM). Leveraging Dolly 2.0 and Langchain, this work demonstrated the potential of advanced language models in our context.

  2. Data Science Research Assistant, O'Neill School of Public and Environmental Affairs

    Aug 2022 — Jan 2023

    Skills Showcased: R, Statistics, Text Analytics, Machine Learning
    I Created an interactive web application for researchers using R (shiny) to extract data from complex CoreLogic housing text datasets based on specified filters and queries, providing easy access to the required data subsets for individual research.

  3. Senior Data Engineer, Quantiphi Inc

    Jul 2019 — Jun 2022

    Skills Showcased: Federated Learning, Model Devgelopment, Data Engineering, Machine Learning, ETL, Tensorflow, PyTorch
    Led a team of five to successfully deliver two projects, namely AppFactory and Federated Learning, in the data engineering track. I also collaborated with Technical Architects to establish a product backlog, catering to Bayer Pharmaceutical as a valued client.
    Engineered a cutting-edge platform for creating Federated Learning experiments, leveraging NVIDIA Clara's client-server model. This innovation enabled our client to train models using local datasets without compromising sensitive medical information.
    Architected and executed a secure cloud-based MLOps platform, known as AppFactory, guiding clients through the entire ML model development process. This encompassed configurable GPU setups, JupyterHub Servers, and essential dependencies such as TensorFlow, PyTorch, and NVIDIA CUDA.
    Successfully achieved cloud-agnostic architecture by reconfiguring 27 Airflow DAGs. Originally operational only on AWS, these DAGs were adapted to seamlessly run on any cloud platform, addressing client needs like Roku.
    Constructed a streamlined data transfer pipeline for clinical trials utilizing Dataflow. This pipeline efficiently retrieved data from SaMD, ensured de-identification via text detection models, and seamlessly loaded the processed data into BigQuery. This optimization resulted in a remarkable 75% reduction in processing time. Furthermore, I crafted an insightful Tableau dashboard to present data-driven insights.
    Designed and implemented a powerful data migration tool called 'Qinetic'. Leveraging Dataproc, Dataflow, and Data Fusion, this tool facilitated the smooth transition of terabytes of data from OLTP sources (Oracle, SQL Server, PostgreSQL) to robust Data warehouses (Redshift, BigQuery), catering to Quantiphi's internal use cases.

My skills

  • Machine Learning
    80%
  • Deep Learning, Natural Language Processing
    75%
  • Data Engineering
    90%
  • MLOps
    90%
  • Databases (Sql, NoSql)
    95%
  • Python, SQL
    90%
  • Behavioural Skills (Leadership, Mentoring, Communication)
    90%

Certifications

  1. Fundamentals of Accelerated Computing with CUDA Python (NVIDIA)

    Completed Feb 2024

    Course Link

  2. Machine Learning Specialization (DeepLearning.AI)

    Completed Sept 2023

    Course Link

  3. GCP Certified Professional Machine Learning Engineer

    Issued Aug 2023

    Verification Link

  4. AWS Certified Solutions Architect Associate

    Issued Nov 2019

    Verification Link

  5. Google Cloud Certified Associate Cloud Engineer

    Issued Oct 2019

    Verification Link

Portfolio

Contact

" width="400" height="300" loading="lazy">

Contact Form