Rithindatta Gundu

About

Built a production LLM-powered fraud detection platform at Wells Fargo, combining OpenAI/Hugging Face models with RAG-based explanations to make flagged transactions interpretable for risk and compliance teams. Delivered low-latency, real-time inference at high scale on AWS (SageMaker + EKS), with strong observability and security controls, reducing manual reviews and false positives in a regulated environment.

Key Strengths

• Built and deployed production LLM-powered fraud detection system in financial services
• Scaled real-time, low-latency inference for high transaction volumes (hundreds of thousands of transactions)
• Reduced manual fraud review effort and lowered false positives
• Strong AWS inference optimization (SageMaker tuning, autoscaling on real metrics, batching, caching, profiling)
• Kubernetes-based microservice orchestration for AI/LLM services (EKS, HPA, blue/green and rolling deploys)
• Security/compliance-first engineering (IAM, VPC isolation, encrypted data flows, secrets management)
• End-to-end observability and performance management (Prometheus/Grafana, SLAs, alert thresholds)
• Modular AI workflow design with measurable success metrics and offline/production evaluation loops
• Effective collaboration with non-technical risk and compliance stakeholders to deliver interpretable outputs

Reference Highlights

• Takes ownership from design through deployment
• Delivers production-grade solutions
• Strong work ethic
• Calm under constraints
• Iterates effectively based on feedback
• Proactive in clarifying requirements early to avoid rework
• Balances security/compliance with low-latency performance needs
• Clear, structured communicator
• Tailors explanations to the audience
• Translates complex AI concepts into practical implications
• Collaborates effectively across teams
• Approachable and reliable
• Builds trust and drives adoption with non-technical stakeholders

Experience

Education

Skills

Languages

Memberships

• ACM