Mohan Shri Harsha Guntu

About

AI/ML practitioner with Northern Trust experience who has shipped production LLM systems (internal support assistant) using RAG, vector databases, orchestration (LangChain/custom pipelines), and rigorous monitoring/feedback loops. Also built AI-driven fraud detection/risk monitoring solutions in a regulated financial environment, emphasizing explainability (SHAP), audit readiness, and stakeholder trust through dashboards and clear communication.

Experience

Education

Key Strengths

• Built and deployed a production LLM internal support assistant using RAG + vector DB
• Improved response time and reduced manual support effort via latency optimizations (prompt reduction, caching, limiting retrieved docs)
• Reliability/quality focus: citation-based answers, content checks, monitoring, and iteration from real user feedback
• Experience orchestrating LLM workflows (LangChain + custom pipelines) for prompt flows, tool calling, RAG, and feedback logic
• Strong evaluation mindset for agents: explicit KPIs (success rate, latency, cost, hallucination, escalation) plus unit/integration testing and continuous monitoring
• Regulated finance ML experience balancing accuracy with explainability and auditability (SHAP, false-positive/recall tradeoffs, temporal consistency)
• Effective collaboration with non-technical risk/compliance stakeholders using dashboards and plain-language explanations to build trust for audits