AI Engineer LLM + Natural Language To Query POC

Employment Type:Short-Term Contract (24 Weeks)Fully Remote (Work from Anywhere)Compensation: 1 Lakh (Fixed Project Fee)About the Role:We are seeking an AI Engineer with strong expertise in Large Language Model (LLM) APIs to develop a Proof of Concept (POC) that converts natural-language questions into accurate SQL or NoSQL queries.A sample database schema (10+ inter-related tables) will be provided by our team. The engineers responsibility is to build a working POC that can interpret plain English queries, understand schema relationships, and generate correct, executable queries with more than 95% accuracy.The output of this engagement will include a demo, a Lucidchart architecture diagram, and detailed technical documentation that our internal team can use to scale the solution further.Key Responsibilities:Develop a self-contained POC that translates natural-language inputs into accurate SQL/NoSQL queries based on the provided schema.Integrate and experiment with multiple LLM APIs (OpenAI GPT-4/5, Anthropic Claude, Google Gemini, Mistral, Cohere, AWS Bedrock).Design prompt structures, validation layers, and accuracy scoring mechanisms to ensure >95% query precision.Evaluate when to use document-based systems (MongoDB, Firestore, Elasticsearch) alongside relational models for complex use cases.Document hybrid storage logicwhen to store data in structured tables vs. document databases.Utilize frameworks such as LangChain, LlamaIndex, Semantic Kernel, Dust, or Amazon Q for model orchestration.Deliver clear Lucidchart workflows showing query generation, context retrieval, and accuracy validation.Provide a Word/PDF technical document explaining the architecture, prompt flow, evaluation metrics, and recommendations.Conduct a final demo and walkthrough session with the internal team.Expected Deliverables:Working POC ApplicationUses the provided schema (10+ tables).Converts natural-language input into correct SQL/NoSQL queries with >95% accuracy.Lucidchart WorkflowVisual representation of the complete pipeline: input prompt LLM query validation output.Technical Documentation (Word/PDF)Setup guide, architecture notes, model design, accuracy testing method, and scale-up roadmap.Demo & Knowledge-Transfer SessionOne live session to explain design rationale and system performance.Required Technical Expertise:Strong proficiency with LLM APIs (OpenAI, Anthropic, Gemini, Cohere, Mistral, Bedrock).Experience in Python (preferred) or for building LLM pipelines.Excellent understanding of SQL, joins, and complex schema relationships.Familiarity with MongoDB or document databases for alternative query structures.Knowledge of LangChain, LlamaIndex, or Semantic Kernel.Understanding of RAG (Retrieval-Augmented Generation), embeddings, and context window optimization.Experience using Lucidchart / Flowchart Maker & Online Diagram Software / Miro for documentation.Strong focus on accuracy, explainability, and reproducibility of AI outputs.Ideal Candidate Profile:48 years total experience, with 1+ years in LLM/NLP/AI projects.Proven experience building natural language to SQL or data query systems.Independent, detail-oriented, and comfortable in a short, fixed-timeline remote engagement.Strong communication and documentation skills for technical hand-offs.Keywords (for Naukri SEO):LLM Engineer, NLP Engineer, Natural Language to SQL, Python AI Developer, LangChain Developer, LlamaIndex, Amazon Q, MongoDB, GPT-4, Claude, Gemini, Bedrock, RAG Engineer, AI Query System, Natural Language Query POC, Generative AI Developer, Prompt Engineer, Remote AI Job, Freelance AI Project, Proof of Concept Engineer.