Opper AI AI Engineer
May 2025 → PresentStockholm
- Develop and maintain the Opper API, powering scalable LLM workflows.
- Built Agent SDKs in Python and TypeScript for production LLM applications.
- Developed the AI Roundtable: multi-model deliberation as a product.
- Designed Taskbench, a real-world LLM benchmarking framework: curated evaluation datasets, a model-agnostic engine (plug-and-play via model string), and a library of evaluator types.
- Built Observer, a context-engineering agent that provides personalised observability across LLM interactions: auto-detects task intent, selects scorers, and improves calls via automated and user-driven in-context learning.
- Deployed autonomous agents in production environments.