Author

Ingeborg Rasmussen

Contributor

4 posts

Covers LLM agent failures via incident-postmortem style: timestamps, decision trees, root causes from raw traces, not dashboards.

Building a Reliable Harness for Multi-Agent AI Systems

Automation software

An agentic harness provides the structural framework necessary to manage communication, error handling, and state persistence between autonomous models.

Ingeborg RasmussenNov 12, 202613 min read

How to Evaluate Telemetry.dev LLM Observability

Automation software

Telemetry.dev LLM observability logs interactions between applications and language models for debugging. It's new with limited independent reviews.

Ingeborg RasmussenOct 31, 202613 min read

Mistral Voxtral 3: Speech Models, Pricing & Benchmarks

Automation software

Mistral Voxtral 3 enables speech-to-text, text-to-speech, and live conversations in automation workflows. It is available via API or self-deployment.

Ingeborg RasmussenOct 30, 202610 min read

ElevenLabs Realtime Speech to Text for Voice Agents

Automation software

ElevenLabs realtime voice agents process speech instantly. Evaluate the latency and accuracy against your application's requirements.

Ingeborg RasmussenOct 27, 202613 min read