Intro

Welcome to my corner of the internet! I'm Ranjeet Nagarkar, a passionate Data Scientist based in San Francisco. With a Master's in Data Science from the University of San Francisco and a strong foundation from the prestigious Indian Institute of Technology Madras, I thrive on solving complex problems and turning data into actionable insights. Throughout my career, I've had the privilege of working on diverse projects, from predictive maintenance applications to recommendation systems and beyond. Whether it's optimizing processes, designing scalable solutions, or harnessing the power of machine learning, I'm driven by a relentless pursuit of excellence. Explore my portfolio to delve into my world of data-driven innovation, featuring projects like the LLM based Twitter complaint resolution project and the Medium recommendation system end-to-end. Feel free to connect with me on LinkedIn. Let's collaborate and embark on exciting data journeys together!

Experience

UnitedHealth Group, USA

Machine Learning Engineer

Designed a scalable chatbot framework using DAG architecture and langchain agents library with Voyager + ReACT prompting. Estimated cost savings of $300,000 with enhanced dialogue management for ReAct chatbot products.

Fine-tuned LLM using PEFT and LoRA techniques for Customer Support Automation tasks, improving KPI by 8%. Leveraged Hugging Face Transformers (BERT, GPT, ALBERT, CLIP, BART) for various NLP tasks.

Deployed the LLM application on cloud platforms including AWS (SageMaker, EC2, EMR, S3, Lambda serverless) and GCP. Enhanced monitoring and operational excellence of LLM systems by leveraging AWS CloudWatch, resulting in a 15% decrease in issue resolution times and a 25% increase in uptime for critical ML and LLM Application workflows.

Conducted A/B testing and causal inference experiments to optimize web applications. Improved neural network architectures, including Convolutional Neural Networks (CNNs) and Transformers, contributing to a 15% boost in model performance and adaptability.



GyanData | Data Scientist | Chennai, India

Indian Navy ILMSAir: Inventory Management System

Analyzed ~20M rows of Data on Snowflake using Apache Spark in Databricks to extract insights on daily equipment procurement.
Optimized procurement process saving an estimated INR 2 million and identified potential logistical bottlenecks worth INR 10 million.
Employed K-means hierarchical clustering to group equipment with akin consumption patterns reducing maintenance cost by ~7-8%.
Deployed Gradient Boosting Regressor, engineered 20+ lagged variables, multi-step time series forecasting reducing procurement by ~5%.

QBSS insurance: Insurance process expedition

Performed POC for Named Entity Recognition(NER) using BioclinicalBERT for medical reports for insurance claim processing.
Communicated with medical experts to incorporate Domain rules into the model improving F1 score from 66% to 85+%.



Indian Institute of Technology Chennai | RBCDSAI

Robert Bosch RBCDSAI: Research on predictive maintenance

Built MATLAB Application for Incremental PCA early fault detection and process monitoring with savings of 9% in operating costs.
Implemented fault detection techniques for error-in-variable sensor data and real-time process monitoring using Datafeed Toolbox.



Personal Projects

LLM based Twitter complaint resolution project

Deployed end-to-end production of LLM-based Twitter complaint resolution project utilizing LLMOps tools, Amazon Kinesis for tweet streaming, LSTM for sentiment analysis, and TwHIN-BERT model for complaint categorization. Developed DAG architecture chatbot application using LLM agents from LlamaIndex library, employed PGvector for company data storage, and AWS deployment.

For more details, visit the Medium blog post here.
The GitHub repository for this project can be found here.


Medium recommendation system end-to-end

Implemented end-to-end Medium recommendation system utilizing MLOps tools, including data scraping from Medium API, data storage in Google bucket, and collections creation in MongoDB using Airflow Composer. Developed Two-tower model-based recommender system for article recommendations to users, alongside a content-based filtering recommender for suggesting similar articles.

For more details, visit the Medium blog post here.
The GitHub repository for this project can be found here.


MATLAB predictive maintenance Application

Created application for analyzing Error-in-Variable time series sensor data from chemical reactors using DIPCA.

The GitHub repository for this project can be found here.


Elements

Text

This is bold and this is strong. This is italic and this is emphasized. This is superscript text and this is subscript text. This is underlined and this is code: for (;;) { ... }. Finally, this is a link.

Heading Level 2

Heading Level 3

Heading Level 4

Heading Level 5
Heading Level 6

Blockquote

Fringilla nisl. Donec accumsan interdum nisi, quis tincidunt felis sagittis eget tempus euismod. Vestibulum ante ipsum primis in faucibus vestibulum. Blandit adipiscing eu felis iaculis volutpat ac adipiscing accumsan faucibus. Vestibulum ante ipsum primis in faucibus lorem ipsum dolor sit amet nullam adipiscing eu felis.

Preformatted

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';

Lists

Unordered

  • Dolor pulvinar etiam.
  • Sagittis adipiscing.
  • Felis enim feugiat.

Alternate

  • Dolor pulvinar etiam.
  • Sagittis adipiscing.
  • Felis enim feugiat.

Ordered

  1. Dolor pulvinar etiam.
  2. Etiam vel felis viverra.
  3. Felis enim feugiat.
  4. Dolor pulvinar etiam.
  5. Etiam vel felis lorem.
  6. Felis enim et feugiat.

Icons

Actions

Table

Default

Name Description Price
Item One Ante turpis integer aliquet porttitor. 29.99
Item Two Vis ac commodo adipiscing arcu aliquet. 19.99
Item Three Morbi faucibus arcu accumsan lorem. 29.99
Item Four Vitae integer tempus condimentum. 19.99
Item Five Ante turpis integer aliquet porttitor. 29.99
100.00

Alternate

Name Description Price
Item One Ante turpis integer aliquet porttitor. 29.99
Item Two Vis ac commodo adipiscing arcu aliquet. 19.99
Item Three Morbi faucibus arcu accumsan lorem. 29.99
Item Four Vitae integer tempus condimentum. 19.99
Item Five Ante turpis integer aliquet porttitor. 29.99
100.00

Buttons

  • Disabled
  • Disabled

Form