Evals in AI: A Deep Dive
Official Schedule Context
- Date/time: 2026-06-29 · 12:10pm-1:10pm
- Track/room: Workshops Day 1 · Track 1
- Speaker(s): Tejas Kumar
- Session type/status: workshop · confirmed
Official Description
“Our evals pass and our velocity is up, so it works.” It’s the most reassuring sentence in AI
engineering and also the most dangerous. Teams are shipping more code than ever while incidents per
PR and change-failure rates climb, and the instruments meant to catch this are quietly broken. This
talk takes apart both halves of that false comfort. First, why velocity lies: the same AI-driven
throughput that lights up your dashboard is what’s eroding quality underneath it. Then we explore
four ways offline evals deceive you: LLM-as-judge bias (your grader rewards confident, wordy, wrong
answers over terse correct ones), staleness, distribution shift between your golden set and real
traffic, and single-score evals that hide which step of an agent actually failed. The centerpiece is
a live demo. We’ll wire up an LLM judge on stage and watch it crown a confident, friendly, factually
wrong answer. Then we’ll fix it live on stage with a three-line rubric change. Same model, different
instrument. From there we’ll build up what to measure instead: traces and spans, production
observability, probe-based evaluation, error budgets, and quality leading indicators that sit beside
every velocity number. Attendees will leave with a five-line checklist they can apply Monday. No
prior eval tooling required. If you’ve ever shipped something agentic and had a nagging feeling the
dashboards were too kind, this is for you.
Related YouTube Video
Harnesses in AI: A Deep Dive — Tejas Kumar, IBM (speaker-match related prior/adjacent AI Engineer video; captions: English auto-captions).
Transcript Status
Related video transcript availability: English auto-captions. Treat this as supporting context, not a recording of this exact scheduled session unless later confirmed. Not fetched yet.
People
Notes
- Pending transcript synthesis when an official recording or confirmed matching video is available.
Supporting Slides
- youtube C_GG5g38vLU slides — extracted from the related public AI Engineer video.
Slide Evidence
- Slide-only cropped deck: youtube C_GG5g38vLU dense slides (1 viable slide images).
- Related slide/OCR pages:
- youtube C_GG5g38vLU dense slides
- youtube C_GG5g38vLU reconstructed slides
- youtube C_GG5g38vLU slides
- Slide-derived terms:
model,const,news,engineering,future,harness,task,hacker,upvote,import,story,session,europe,lght,https,news.ycombinator.com,stories,highest-ranked