NGR
Work Experience

4+ years
of impact

BNY
Oct 2024 — Current
Illinois

Data Scientist

BNY·Full-time
2M+ requests
0.5M+ Transactions
40% workload reduction
35% error reduction
  • Engineered real-time GenAI fraud detection using async LLM calls via AWS Bedrock and Claude APIs with rate-limiting and retry logic, achieving 84% precision.
  • Built high-performance FastAPI RESTful backend microservices on AWS Lambda, serving 2M monthly requests and reducing transaction errors by 35%.
  • Architected multi-agent GenAI workflows with LangChain tool-calling and reasoning chains for automated customer classification and transaction monitoring.
  • Engineered LLM-based document classification achieving 82% accuracy on eligibility predictions, reducing manual review workload by 40% in production.
Vivma
Mar 2021 — Jul 2023
India

Data Scientist

Vivma Software Inc·Full-time
25% engagement boost
79% AUC-ROC
28% cost reduction
30% faster decisions
  • Increased user engagement by 25% by constructing machine learning ensemble models including scikit-learn, XGBoost, and Random Forest.
  • Built intent classification system using BERT transformers and spaCy NLP chatbot across 50 intent categories, reducing response times from hours to minutes.
  • Deployed production churn prediction system leveraging Apache Airflow and Azure Synapse for data pipeline orchestration, achieving 79% AUC-ROC.
  • Developed ARIMA and LSTM time-series forecasting models on Azure Functions, reducing inventory costs by 28% with 82% MAPE.
OPEN TO NEW OPPORTUNITIES · REMOTE · OPEN TO RELOCATE
Highlights
84%
GenAI fraud detection precision
2M+
Monthly API requests served
40%
Manual review workload reduced
25%
Increase in user engagement
28%
Inventory costs reduced
30%
Faster feature launch decisions