Cognisol logo
CognisolPosted 1 month ago

AI Architect

On-siteCarlsbad, California, United States

ContractSmall

Job Summary

Architect end-to-end LLM solutions for chatbot applications, semantic search, and domain-specific assistants using Databricks Unity Catalog for centralized governance, metadata management, and audit trails. Design modular, scalable workflows including prompt orchestration, RAG, and real-time inference pipelines while integrating feedback loops to continuously refine model performance. Optimize latency and cost through fine-tuning and token management, overseeing secure production deployments with strict access controls and compliance alignment. Lead teams on data quality and responsible AI practices across enterprise-scale use cases.

Required Qualifications

  • Currently based in CA
  • 7+ years in AI/ML solution architecture
  • 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)

Desired 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

Hiring someone like this?

Get your role in front of qualified candidates on Sorce.

Get started

Apply to this job in one click with Sorce

Apply on Sorce