How this entry appears when shared on social media
Small, focused experiments with 2 people surfaced which streams to keep vs deprioritise and produced reusable tooling (an LLM wrapper + experimental flag) that should speed future decisions.
Work·Engineering, Process, ML
Under tight contractual obligations and unclear vectors of work, conducted a set of experiments with 2 team members. We identified and prioritised/deprioritised streams of work. Result: we got a tool that wraps everything to an LLM — add-experimental-flag/SKILL.md.
Small, focused experiments with 2 people surfaced which streams to keep vs deprioritise and produced reusable tooling (an LLM wrapper + experimental flag) that should speed future decisions.
Contractual deadlines and unclear work directions created pressure to triage work fast. We ran experiments with 2 team members to test assumptions and surface high-value streams quickly.