Due to project requirements and business needs, we are only considering candidates who are currently based in CA
Description:
We are seeking a seasoned AI Architect with strong experience in Generative AI and Large Language Models (LLMs)including OpenAI, Claude, and Geminito lead the design, orchestration, and deployment of intelligent solutions across complex use cases. You will architect conversational systems, feedback loops, and LLM pipelines with robust data governance, leveraging the Databricks platform and Unity Catalog for enterprise-scale scalability, lineage, and compliance.
Role Scope / Deliverables:
Key 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 trailso 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.
Key Skills:
- 7+ years in AI/ML solution architecture, with 2+ years focused on LLMs and Generative AI.
- Strong experience working with OpenAI (GPT-4/o), Claude, Gemini, and integrating LLM APIs into enterprise systems.
- Proficiency in Databricks, including Unity Catalog, Delta Lake, MLflow, and cluster orchestration.
- Deep understanding of data governance, metadata management, and data lineage in large-scale environments.
- Hands-on experience with chatbot frameworks, LLM orchestration tools (LangChain, LlamaIndex), and vector databases (e.g., FAISS, Weaviate, Pinecone).
- Strong Python development skills, including notebooks, REST APIs, and LLM orchestration pipelines.
- Ability to map business problems to AI solutions, with strong architectural thinking and stakeholder communication.
- Familiarity with feedback loops and continuous learning patterns (e.g., RLHF, user scoring, prompt iteration).
- Experience deploying models in cloud-native and hybrid environments (AWS, Azure, or GCP).
Preferred Qualifications:
- Experience fine-tuning or optimizing open-source LLMs (e.g., LLaMA, Mistral) with tools like LoRA/QLoRA.
- Knowledge of compliance requirements (HIPAA, GDPR, SOC2) in AI systems.
- Prior work building secure, governed LLM applications in highly regulated industries.
- Background in data cataloging, enterprise metadata management, or ML model registries