Technical Focus Areas
Deep dives into the architecture of modern AI platforms
Deep dives into the architecture of modern AI platforms
Cost control is an engineering constraint, not an accounting problem. My work covers embedding financial gates into CI/CD, using DSPy to optimize prompts for smaller models (Llama 8B vs GPT-4), and managing Provisioned Throughput.
View FinOps Articles on the Blog →
Solving the 'Data Bleed' problem in multi-tenant AI. Techniques include Hybrid Search tuning, enforcing Row-Level Security (RLS) at the Vector Search layer, and managing data freshness with Delta Change Data Feed.
View RAG Security Articles on the Blog →
Moving beyond chatbots to 'Digital Workers.' Architectures for Router Agents, integrating structured data (SQL) via Databricks Genie, and managing asynchronous tool execution for high performance.
View RAG Security Articles on the Blog →
Moving beyond chatbots to 'Digital Workers.' Architectures for Router Agents, integrating structured data (SQL) via Databricks Genie, and managing asynchronous tool execution for high performance.
View RAG Security Articles on the Blog →