coming in planning

ai engineer pro

specialist ai systems where the agent layer alone isn't enough. voice agents, multimodal, long-context vs rag, computer use, and on-device inference.

5 lessons|2 modules|~3 hours

what you’ll learn

ship a production voice agent with a real latency budget
design multimodal pipelines that handle screenshots, charts, and image-aware support flows
reason about long-context vs rag and pick the right tool without product-chrome arguments
understand computer-use agents and on-device inference tradeoffs at the architectural level

curriculum

planning sketch

this is a rough curriculum we’re still planning. modules and lessons are likely to shift before any lesson is recorded. want to shape it? mail@karnstack.com.

module one

specialist case studies

~65 min2 lessons

01voice agent: livekit and pipecat referencecoming soon35m

02multimodal pipeline: image and textcoming soon30m

module two

open problems

~90 min3 lessons

03long context vs rag vs hybrid retrievalcoming soon30m

04computer use and browser agentscoming soon30m

05on-device and edge inferencecoming soon30m

frequently asked

when does this launch?: in planning, sequenced after production-agents. the curriculum on this page is a sketch. modules and lessons are likely to shift before any lesson is recorded.
how is this different from production-agents?: course 2 covers the agent layer end-to-end (loops, tools, memory, runtime, sandboxing). course 3 covers specialist systems where that layer alone isn't enough: voice latency budgets, multimodal pipelines, computer use, and on-device tradeoffs.