Sumer Joshi

About

Built a chatbot for a learning management system during a Deep Atlas bootcamp by mapping an end-to-end RAG architecture (document ingestion, Qdrant-based retrieval scoring, and LLM response synthesis). Previously at Rally Health/UnitedHealthcare, diagnosed load-related memory spikes with JMeter and improved stability by migrating caching from Guava to Redis, and also supported adoption through UI A/B testing in a technical marketing engineer rotation.

Experience

Education

Key Strengths

• Designed end-to-end RAG chatbot workflow from user request through retrieval and LLM synthesis
• Implemented retrieval relevance scoring using vector DB scores (Qdrant) to filter/return higher-quality results
• Real-time performance diagnosis using load testing (JMeter) and root-cause analysis
• Resolved production stability issue by replacing Guava Cache with Redis after identifying memory spikes under load
• Adapts technical demos/workshops to audience level; uses demo-first approach with deep Q&A
• Improved product adoption via UI A/B testing and iterative design changes in a healthcare context

Reference Highlights

• Eager to learn
• Takes time to understand documentation and systems
• Collaborates well with senior engineers
• Steady improvement based on feedback
• Proactively identifies potential issues
• Reliable
• Detail-oriented
• Good communicator
• Asks relevant questions and contributes ideas
• Insightful during pairing sessions (peer compliments)
• Adaptable to shifting priorities
• Takes ownership and drives investigations to resolution
• Maintains effective communication with external stakeholders
• Follows up to ensure issues remain resolved