Athwika Gade

About

LLM/RAG engineer at Connex AI who built and deployed a production healthcare agent to extract clinical insights from medical data/notes. Strong focus on real-world reliability—hallucination mitigation (citations, schema validation, confidence thresholds, rejection logic), custom LangChain orchestration (query rewriting, fallback paths), and production evaluation/observability—while collaborating closely with clinical SMEs to ensure clinical fit and time savings.

Experience

Education

Key Strengths

• Built and deployed a production LLM healthcare agent for clinical insight extraction
• Designed end-to-end RAG systems (ETL + retrieval + FastAPI services) for production use
• Hallucination reduction via constrained RAG prompting, citations, schema validation, confidence thresholds, and rejection logic
• Production reliability engineering: retries, fallback models/paths, continuous evaluation, and logging
• Custom LangChain orchestration for inconsistent clinical queries (query rewriting + fallback logic) to improve retrieval relevance and response stability
• Performance optimization balancing latency vs. reliability; reduced error rates and improved response consistency
• Structured testing/evaluation approach using curated datasets, edge cases, regression, and production A/B testing
• Effective collaboration with non-technical clinical SMEs to translate workflows into AI requirements and iterate via prototypes