Vimala Devi

About

ML/AI engineer with hands-on experience building production systems in financial services, including a real-time underwriting analytics platform at Hartford Financial Services. Stands out for combining classic ML, low-latency API deployment, monitoring, and emerging LLM/RAG design patterns, with measurable impact including 20% better decision accuracy, sub-200ms latency, and 5M+ records processed daily.

Key Strengths

• End-to-end ownership of production ML systems
• Real-time ML deployment for financial underwriting
• Built low-latency FastAPI inference services
• Designed monitoring for performance and model drift with Prometheus/Grafana
• Strong experimentation and A/B testing mindset
• RAG-first approach for accuracy and hallucination reduction in financial use cases
• Structured LLM evaluation across groundedness, relevance, benchmarks, and human review
• Scaled systems processing 5M+ records daily
• Delivered sub-200ms real-time inference latency
• Improved decision accuracy by 20%
• Reduced underwriting decision time from days to minutes
• Increased case throughput by 15%
• Translated research LLM prototypes into production-oriented document intelligence pipelines
• Strong Python service/library design for multi-team reuse
• Improved reliability through shared internal libraries, standardized validation, and resilient API adapters

Experience

Education

Skills

Languages