Glean
AI / ML EngineerAppliedAug 6, 2026

Deduplicate operational-answer evidence while preserving updates

More evidence does not always make an answer more useful.

Repeated updates can look like corroboration while crowding out the change that matters.

I don’t need the same update three times—I need to know what changed last.

Stefano Gomes · Enterprise Reliability Engineer

Uses cross-system retrieval during live incident response and post-incident follow-up.

What pulls against what

  • context efficiency vs. update fidelity
  • semantic similarity vs. source lineage
  • latency recovery vs. citation quality
  • compression speed vs. reviewable evidence

What is at stake

Redundant context slows answers and hides distinct updates. The work must improve evidence selection without treating legitimate corroboration as noise

Why Glean

At Glean, the pattern can emerge when the same work is documented across connected systems.

Written for

retrieval ML engineerevaluation-focused applied scientistML systems engineer

This is the setup. The work is inside.

Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.