Pranava Reddy Kothapally

About

LLM/AI engineer who has deployed production RAG conversational analytics and Text-to-SQL systems over Snowflake and curated data marts, emphasizing enterprise-grade guardrails for accuracy, security, and cost. Notable for a structured approach to reducing hallucinations (curated metric/table registry, SQL validation, RBAC, and citation-backed responses) and for building resilient, observable multi-step agent workflows using LangChain/LlamaIndex and Airflow.

Experience

Education

Key Strengths

• Built and deployed production LLM apps for natural-language querying of enterprise data
• Deep expertise mitigating Text-to-SQL hallucinations via grounding, planning/execution separation, and SQL validation gates
• Strong security and governance focus (RBAC, mandatory filters, approved join paths, data leakage prevention, cost controls)
• Designs resilient multi-step LLM workflows with isolation, retries, fallbacks, and graceful degradation
• End-to-end ownership of LLM lifecycle (ingestion, orchestration, evaluation, monitoring, continuous improvement)
• Rigorous testing/evaluation approach (golden sets, unit tests, adversarial cases; tracks groundedness, hallucination rate, latency, cost)
• Effective cross-functional delivery with non-technical stakeholders (finance/ops), aligning on trust/safety requirements and SLAs

Reference Highlights

• Shipped AI/LLM integrations and infrastructure to production
• Led design and launch of an LLM-powered customer support assistant
• Translates vague user needs into clear product requirements
• Balances model capabilities with UX and cost constraints
• Iterates quickly based on real user feedback
• Effective cross-team collaboration
• Explains technical concepts in business-friendly language
• Strong communication skills
• Aligns stakeholders on shared goals
• Clear and concise communicator
• Tailors explanations to the audience
• Breaks down complex ideas into practical terms
• Asks clarifying questions to ensure understanding