Monisha DhanaVijeya

About

Built and shipped a production multi-agent conversational AI platform (Monitor agent + RAG + 4 additional agents) with enterprise REST APIs, using ChromaDB-grounded WCAG knowledge to keep responses accurate while varying tone via personality modes and conversation memory. Has experience at LinkedIn delivering technical demos and pre-sales guidance to both engineering teams and C-level stakeholders, acting as a translator between sales and technical teams to drive adoption.

Experience

Education

Key Strengths

• Took a multi-agent LLM system from proof-of-concept to production with enterprise-ready REST APIs
• Strong stakeholder communication via real-time Streamlit dashboards and clear documentation for non-technical audiences
• Improved conversational quality by reducing generic/repetitive responses using personality modes, conversation memory (last 20 interactions), and prompt engineering
• Built accuracy safeguards by grounding responses in a ChromaDB RAG store with structured guidelines and code samples
• Systematic real-time troubleshooting of LLM/agent pipelines (prompts, logs, integration logic, data quality) with quick fixes plus scalable follow-ups
• Effective technical demoing for developer audiences using code-first walkthroughs, architecture diagrams, and layered explanations across experience levels
• Bridges sales and engineering to support deal cycles—translates technical capabilities into business value/ROI and converts customer pain points into actionable specs
• Led live production CMS backend redesign and migration with phased dual-stack rollout
• Built validation layer to compare legacy vs new pipeline outputs and migrate content types incrementally
• Diagnosed and fixed DB connection pooling misconfiguration; improved latency by 40%
• Handled malformed legacy XML via preprocessing plus automatic fallback routing and logging
• Systematic AWS incident debugging across app/network layers (CloudWatch, NAT gateway limits, TCP timeouts)
• Implemented queue-based rate limiting to eliminate timeouts with minimal performance impact (+0.3s)
• Designed pluggable Python middleware for customer-specific WCAG reporting without disrupting core pipeline
• Improved performance using LRU caching to avoid repeated expensive DOM analysis
• Strong customer-facing communication—guided client to fix conversion pixel placement and explained business impact