Available for opportunities

Jyoshna Pasupuleti

Senior Data Scientist & AI/ML Engineer specializing in LLMs, RAG architectures, and production-grade machine learning systems across retail, healthcare, and insurance.

7+
Years of experience
4
Industries
50%
Avg. efficiency gain
01. about

Building AI that
actually ships.

Hi, I'm Jyoshna — a Senior Data Scientist and AI/ML Engineer with 7+ years of experience building scalable machine learning and Generative AI solutions. I've worked across retail, healthcare, and insurance, turning complex data into real-world production systems.

I specialize in LLMs, RAG architectures, and end-to-end ML pipelines, with deep expertise in Python, TensorFlow, PyTorch, and cloud platforms like AWS and Azure. Currently based in New York, I'm passionate about engineering AI that doesn't just work in notebooks — it works in production.

I hold a Master's in Big Data Analytics & IT from UCM, USA and a Bachelor's in Electronics & Communication Engineering from REVA University, India.

Pipeline efficiency

Reduced ML pipeline processing time by 30–40% through optimized feature engineering and workflow orchestration.

🎯

Model accuracy

Improved model accuracy by 15–25% across NLP and forecasting use cases across multiple domains.

🤖

GenAI impact

Built and deployed scalable LLM-based RAG solutions reducing manual effort by 50% in enterprise workflows.

02. experience

Where I've
worked.

Apr 2025 – Present
Synechron
New York, NY
AI/ML Engineer
  • Built end-to-end ML systems for batch and real-time inference, improving deployment efficiency by ~35%.
  • Fine-tuned BERT and RoBERTa transformer models, improving classification accuracy by 20%.
  • Developed scalable NLP pipelines for large-volume unstructured data processing.
  • Containerized ML workloads with Docker and Kubernetes for fault-tolerant production environments.
  • Implemented monitoring frameworks to detect data drift, reducing production issues by 30%.
PythonTensorFlowPyTorch NLPDockerKubernetes AWSSpark
Jun 2023 – Mar 2025
The Home Depot
Atlanta, GA
Senior Data Scientist
  • Developed large-scale Spark pipelines processing high-volume retail data, cutting processing time by 35%.
  • Built ML models for demand forecasting, improving accuracy by 18% and optimizing inventory planning.
  • Delivered interactive Tableau dashboards for KPI monitoring and data-driven decisions.
  • Optimized SQL queries and workflows, reducing execution time by 40%.
  • Built and deployed AWS-based data pipelines using S3, Lambda, and Redshift.
PythonSQLSpark AWSTableauDocker
Aug 2020 – Jul 2022
Max Healthcare
New Delhi, India
Data Scientist
  • Built scalable data pipelines using Spark and Azure Data Factory, improving efficiency by 30%.
  • Enabled near real-time analytics via Azure Event Hubs and Stream Analytics integration.
  • Developed ML models for clinical and operational analytics to support data-driven decisions.
  • Created Power BI dashboards with role-based access for diverse stakeholders.
PythonSQLAzure SparkPower BITensorFlow
Jun 2018 – Jul 2020
TATA AIG Insurance
Mumbai, India
Data Analyst
  • Automated reporting workflows, reducing manual effort by 50% and improving turnaround time.
  • Developed regression and classification models to support risk analysis and business decisions.
  • Built Tableau dashboards to visualize business performance and key metrics.
  • Managed AWS-based batch processing workflows to ensure reliable data processing.
PythonSQLR TableauAWS
03. projects

GenAI & ML
work.

01 /

RAG-Based Intelligent Document Search

Built an end-to-end intelligent document search system using LangChain and vector databases, enabling semantic search across enterprise documents.

↓ 60% manual lookup time
LangChainVector DBsRAGPython
02 /

LLM-Powered Enterprise Chatbot

Developed an enterprise chatbot using prompt engineering and embeddings, improving response accuracy and user engagement for internal teams.

↑ Response accuracy & engagement
LLMsPrompt EngineeringEmbeddings
03 /

Scalable Document Ingestion Pipeline

Designed an end-to-end pipeline for document ingestion, chunking, embedding, and retrieval to power scalable AI applications in production.

Production-grade at scale
LangGraphAWSPythonDocker
04 /

Retail Demand Forecasting System

Built ML models incorporating seasonality, promotions, and regional trends for demand forecasting at The Home Depot, optimizing inventory planning.

↑ 18% forecast accuracy
SparkPythonAWSTableau
05 /

Transformer NLP Classification Pipeline

Fine-tuned BERT and RoBERTa models for text classification on large-scale unstructured data with full MLOps monitoring and retraining automation.

↑ 20% classification accuracy
BERTRoBERTaPyTorchMLflow
06 /

Healthcare Real-Time Analytics Platform

Enabled near real-time clinical and operational analytics at Max Healthcare using Azure Event Hubs and ML-powered dashboards for stakeholder insights.

↑ 30% processing efficiency
AzureSparkPower BIPython
04. skills

Tech stack &
tools.

Generative AI & LLMs
RAG LangChain LangGraph Prompt Engineering Vector DBs Embeddings
ML & Deep Learning
TensorFlow PyTorch Transformers BERT NLP Regression Classification
Cloud & MLOps
AWS Azure SageMaker Docker Kubernetes MLflow CI/CD
Data & Big Data
Spark Kafka Hadoop SQL Redshift
Languages
Python SQL R
Visualization
Tableau Power BI
05. contact

Let's build something
together.

I'm open to new opportunities, collaborations, and interesting AI/ML projects. Feel free to reach out.