Vetted Generative AI Engineers

“ML/NLP practitioner who built a retrieval-augmented generation (RAG) system for large financial and operational document sets using Sentence-Transformers (all-mpnet-base-v2) and a vector DB (e.g., Pinecone), with a strong focus on retrieval evaluation and chunking strategy optimization. Experienced in entity resolution (rules + embedding similarity with type-specific thresholds) and in productionizing scalable Python data workflows using Airflow/Dagster and Spark.”

“GenAI engineer who has deployed production LLM/RAG chatbots for internal document search, focusing on reliability (hallucination reduction via prompt guardrails + retrieval filtering) and performance (latency improvements via caching). Experienced with LangChain/LangGraph orchestration for multi-step agent workflows and iterates using monitoring/logs and benchmark-driven evaluation while partnering closely with product and business teams.”

“Built and shipped a production RAG-based enterprise knowledge assistant to replace slow/inaccurate search across millions of documents, using LangChain orchestration with GPT-4/LLaMA and vector databases. Strong focus on production constraints—latency, hallucination control, and cost—using hybrid retrieval, guardrails, LLM-as-judge validation, and model routing, and has experience translating non-technical stakeholder pain points into measurable outcomes.”

“Healthcare-focused LLM engineer who deployed a production triage and clinical knowledge retrieval assistant using RAG and LangGraph-orchestrated multi-agent workflows. Emphasizes clinical safety and compliance with robust hallucination controls, HIPAA/PHI protections (tokenization, encryption, audit logging, zero-retention), and human-in-the-loop escalation; reports a 75% latency reduction in a healthcare agent system.”

“Built a production multi-agent orchestration platform to automate healthcare claims and HR workflows, combining LangChain/CrewAI/AutoGPT with RAG (FAISS/Pinecone) and fine-tuned open-source LLMs (LLaMA/Mistral/Falcon) in private Azure ML environments to meet HIPAA requirements. Emphasizes rigorous agent evaluation/observability (trajectory eval, adversarial testing, LLM-as-judge, drift monitoring) and reports measurable outcomes including 35% faster claims processing and 40% fewer chatbot errors.”

“Built and deployed a production LLM-powered workflow assistant that automated internal marketing/production business tasks (document summarization, repeated Q&A, status updates). Demonstrates end-to-end applied LLM engineering: modular RAG architecture, hallucination/latency mitigation, automated evals to prevent prompt regressions, and Azure-based orchestration (Functions/Logic Apps) with monitoring and controlled rollouts.”

“Designed and deployed a production LLM agent platform at the National Institutes of Health to reduce time spent searching fragmented internal documentation, combining RAG grounding with multi-step tool-calling workflows and integration into legacy services via inference APIs. Emphasizes production-grade reliability through automated evaluation on real queries, guardrails/safe-failure behaviors, and ongoing A/B testing and monitoring, and has experience translating non-technical stakeholder goals into measurable success metrics.”

“AI/LLM engineer with healthcare domain experience who built a production clinical support “chart bot” for Molina, including PHI-safe ingestion of 200k+ PDF policies, vector retrieval, and a fine-tuned LLaMA served via vLLM on ECS Fargate. Demonstrated measurable performance wins (HNSW + namespace partitioning; 30% inference latency reduction) and a rigorous evaluation/monitoring approach, while partnering closely with nurses and operations teams to shape workflows and guardrails.”

“Backend engineer who built and owned an AI-powered financial research product end-to-end, using a typed NestJS/GraphQL backend with LangGraph-style agent routing to produce sourced, structured financial analysis. Emphasizes finance-grade correctness (Zod validation, metric registries, unit/empty-result guardrails) while keeping latency low via batching, caching, and fast token streaming, and has led incremental migrations using strangler/feature-flag/shadow traffic patterns.”