AI ARCHITECT

Cognisol is seeking a seasoned AI Architect with strong experience in Generative AI and Large Language Models (LLMs). The role involves leading the design, orchestration, and deployment of intelligent solutions, including architecting conversational systems and optimizing model performance.

Responsibilities:

• Architect end-to-end LLM solutions for chatbot applications, semantic search, summarization, and domain-specific assistants
• Design modular, scalable LLM workflows including prompt orchestration, RAG (retrieval-augmented generation), vector store integration, and real-time inference pipelines
• Leverage Databricks Unity Catalog for: centralized governance of AI training and inference datasets, managing metadata, lineage, access controls, and audit trails, cataloging feature tables, vector embeddings, and model artifacts
• Collaborate with data engineers and platform teams to ingest, transform, and catalog datasets used for fine-tuning and prompt optimization
• Integrate feedback loop systems (e.g., user input, signal-driven reinforcement, RLHF) to continuously refine LLM performance
• Optimize model performance, latency, and cost using a combination of fine-tuning, prompt engineering, model selection, and token usage management
• Oversee secure deployment of models in production, including access control, auditability, and compliance alignment via Unity Catalog
• Guide teams on data quality, discoverability, and responsible AI practices in LLM usage