Ideas from the work
Perspectives and practical lessons in ai & ml.
6 articles
Newest first
Three practical tutorials for building a local DGX Spark media workflow with Hermes, ComfyUI, moving video, experimental Urdu voiceover and audio finishing, including...
Lessons from Labverse Agent on durable messages, explicit takeover, grounded answers, and honest delivery states when automation meets customer service.
Agentic applications can break a goal into tasks and select tools within configured limits. Their usefulness depends on how clearly those limits are defined, how tool...
Anthropic's Model Context Protocol has quietly become one of the most consequential standards in the AI tooling ecosystem. MCP defines a universal interface for...
A production retrieval-augmented application needs an authorized knowledge source, suitable retrieval, grounded answers, and an evaluation process. Query decomposition,...
AI applications face risks including prompt injection, unintended data exposure, and unauthorized tool use. Define the data and action boundaries for each application,...