Hello, I'm Rupsa Banerjee

Empowering Human-AI Interaction by Transforming Research into Real-World Impact—Where Speech Meets Intelligence, and Machine Learning Drives the Future of Conversational Systems

Download Resume Schedule Consultation Let's Collaborate

Skills

Programming Languages

Python, SQL, Javascript, HTML/CSS

Development Tools

Git, Jira, Windows, Linux, Docker

Data Visualization

Tableau, Matplotlib, Data Modeling, Data Warehousing

Generative AI & LLMs

GPT, BERT, Claude, Hugging Face, Prompt Engineering, RAG, Vector DB

Big Data & Cloud

Hadoop, Spark, MongoDB, AWS, Azure

Web & API Development

FastAPI, RESTful API, WebSocket, Django, React, WebRTC

Machine Learning & AI

TensorFlow, PyTorch, Scikit-learn, NLP, Computer Vision, Reinforcement Learning

MLOps & Soft Skills

CI/CD Pipelines, ETL/ELT, Workflow Orchestration, Effective Communication, Leadership, Adaptability

Professional Experience

Independent AI Consultant

Current Position
Duration Feb 2025 - Present
Location New Jersey, USA
Supervisor Dr. Pantelis Monogioudis Dept. of Data Science, New Jersey Institute of Technology
  • Architected production-ready, real-time Conversational AI Platform for cognitive health applications with WhisperLive STT, CUDA acceleration, and multi-agent intent recognition framework
  • Implemented ultra-low latency text-to-speech using HuggingFace Kokoro-82M achieving sub-300ms response times with multi-speaker voicepacks and voice blending capabilities
  • Fine-tuned and deployed Sesame CSM model with LoRA for semantic/acoustic token modeling, integrated with RAG-based contextual memory via Model Context Protocol (MCP)
  • Deployed scalable system using FastAPI, WebSocket streaming, NiceGUI frontend with WebRTC, containerized with Docker/UV, and managed via Loguru/Pydantic configuration

Machine Learning Researcher

Research Position
Duration Jan 2024 - Jan 2025
Institution New Jersey Institute of Technology
Supervisor Dr. Mengjia Xu Dept. of Data Science, New Jersey Institute of Technology
Advisor Dimitrios Pantazis Director of MEG Core Neuroimaging, Massachusetts Institute of Technology
  • Developed a TransformerG2G-based framework for analyzing dynamic brain connectivity using MRI data, in collaboration with teams from MIT and Oxford
  • Leveraged self-attention mechanisms and uncertainty modeling to improve link prediction accuracy (novelty score: 0.6), achieving significant gains in MAP and MRR
  • The model effectively captured temporal brain dynamics, enabling early diagnosis of neurological disorders and contributing to impactful public health outcomes

Search Engine Optimization Data Analyst

Analytics Position
Duration Jan 2018 - May 2020
Company Irvine Infocom Pvt. Ltd. | India
  • Optimized large-scale ETL pipelines across AWS, Azure, and GCP to process data from Google Search Console and SEMrush
  • Applied machine learning techniques for keyword ranking, content clustering, and backlink strategy, resulting in improved search visibility and automated SEO workflows
  • Utilized Tableau and Excel for advanced data visualization, combined with statistical analysis and Git-based version control to deliver data-driven insights and enhance content performance

Data Driven Web Optimization Analyst

Freelancing
Duration June 2020 - Dec 2021
Type Freelancing | Online Collaborations
  • Enhanced Web Optimization through freelancing and online collaborations, implementing data-driven strategies for website visibility
  • Specialized in keyword optimization and digital marketing analytics to improve client web presence and performance metrics

Academic Projects

Human Action Recognition

Advanced deep learning model for real-time human action classification using computer vision techniques

Python Deep Learning Computer Vision
View Code

Azure Data Engineering Project - Tokyo Olympics

Comprehensive data engineering pipeline for Tokyo Olympics data using Azure cloud services and ETL processes

Azure Data Engineering ETL
View Code

Twitter Sentiment Analysis ETL Project

End-to-end ETL pipeline for analyzing Twitter sentiment using natural language processing and machine learning

Python NLP Sentiment Analysis
View Code

Drone Detection Using YOLOv8

Advanced machine learning project for real-time drone detection using YOLOv8 object detection algorithms

YOLOv8 Object Detection Computer Vision
View Code

Stock Price Prediction with RNN

Recurrent Neural Network implementation for predicting stock prices using time series analysis and deep learning

RNN Time Series Financial Analysis
View Code

Walmart Sales Data Analysis and Ad-Hoc Query Project

Comprehensive retail analytics project with SQL-based ad-hoc queries and data visualization for sales performance insights

SQL Data Analysis Retail Analytics
View Code

About me

Rupsa Banerjee

Data Science & AI Specialist

Transforming Research into Real-World Impact

Data scientist and AI researcher specializing in machine learning and conversational AI systems.

Expert in developing real-time AI platforms and analyzing brain connectivity using deep learning.

Passionate about transforming complex data into actionable insights and innovative solutions.

Contact info

Phone

+616-290-2999

Email

rupsabanerjee01@gmail.com

LinkedIn

Connect with me

Adress

New Jersey-USA, zip:07034