Adarsha Nagaraja

Machine Learning Engineer

Plaid

Jul 2025 - Current (1 year 4 months)

• Engineered LendSmart AI Platform, a scalable end-to-end system for credit risk scoring and lending insights, analyzing millions of banking
transactions daily, improving loan approval accuracy by 32% while reducing default risk and operational latency.
• Built robust data pipelines with Python, SQLAlchemy, and Pydantic, ingesting multi-institution financial datasets into PostgreSQL and feature stores
with 99.7% schema validation, ensuring consistent, high-quality inputs for AI-driven credit scoring models and internal APIs.
• Developed NLP-based risk modeling with FinBERT and spaCy to extract transaction patterns, recurring behaviors, and income signals, improving
credit score precision by 20%, supporting adaptive

Machine Learning Engineer

Tietoevry

Jul 2023 - Jan 2024 (7 months)

Enterprise GenAI platform automating contract analysis and compliance extraction from unstructured legal documents,
reducing manual review time by 65% while processing 150K+ documents annually across procurement, legal, and risk
management teams.
• Reduced contract review time by 65% (from 180 to 63 minutes per contract) as measured by legal team productivity
metrics, by architecting and deploying a RAG-powered GenAI platform that automatically extracts key clauses,
obligations, and risk factors from complex legal documents
• Improved clause extraction accuracy by 42% as measured by legal validation reviews, by implementing hybrid retrieval
system combining dense embeddings (sentence-transformers) and spar

Data Scientist

Tietoevry

Oct 2020 - Jun 2023 (2 years 9 months)

End-to-end predictive analytics platform identifying at-risk customers and optimizing retention strategies across 2M+
subscriber base, reducing churn rate by 23% and generating $4.2M in retained annual recurring revenue through targeted
interventions.
• Reduced customer churn by 23% (from 8.7% to 6.7%) as measured by quarterly retention metrics, by designing and
deploying gradient boosting–based churn prediction model with 0.88 ROC-AUC achieving 78% precision in identifying
high-risk customers 30 days before cancellation, leveraging strong analytical skills to inform feature selection and model
evaluation
• Generated $4.2M in retained annual revenue as measured by prevented cancellations, by building autom

Associate Data Scientist

Tietoevry

May 2018 - Sep 2020 (2 years 5 months)

Comprehensive analytics solution tracking 8M+ shipments annually across global supply chain network, enabling data-driven
logistics optimization that reduced delivery delays by 31% and improved inventory turnover by 18% through predictive
insights.
• Reduced delivery delays by 31% (from 12.4% to 8.6% late shipments) as measured by on-time delivery metrics, by
building comprehensive supply chain analytics platform identifying bottlenecks in logistics operations including carrier
performance, warehouse processing times, and route optimization opportunities
• Improved inventory turnover by 18% as measured by days-inventory-outstanding, by developing SQL-based analytical
models in BigQuery tracking 8M+ annual

Machine Learning Research Intern

Indian Institute of Information Technology Allahabad

Jan 2018 - Mar 2018 (3 months)

Translated complex business problems into analytics-driven solutions serving 10,000+ users as measured by adoption metrics, by analyzing user
behavior data, identifying pain points through statistical analysis, designing ETL pipelines for feature extraction, and implementing ML-powered
recommendations achieving 99.9% system reliability
• Accomplished 40% improvement in system performance as measured by query benchmarks, by conducting query optimization and
troubleshooting of 200+ database queries, identifying bottlenecks through execution plan analysis, implementing data-driven optimization
strategies, and validating results through A/B testing with confidence intervals
• Mentored technical team of 8 engineers

About

Detail

Timeline

Résumé

Adarsha Nagaraja

About

Detail

Behavioral traits