Slides: The Production AI Playbook: Deploying Agents at Enterprise Scale — Sandipan Bhaumik, Databricks

Source Video

The Production AI Playbook: Deploying Agents at Enterprise Scale — Sandipan Bhaumik, Databricks

Relationship To World's Fair 2026

These slides are extracted from a public AI Engineer YouTube video connected to World's Fair 2026. Speaker-matched clips are supporting context unless later confirmed as exact session recordings; official livestream recordings are day-level/event-level source material.

Related Scheduled Sessions

No individual scheduled session mapping has been assigned yet; treat this as an event livestream deck.

Extracted Slides

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.97
Text source: advanced OCR rapidocr-live/bright-screen/contrast.
OCR decision: ready — small timeline labels and caption text across a multi-element slide

Slide text:

THEPROBLEM

The pattern you already know

AIE Weeks1-4 Weeks4-8 Weeks8-12 Week14 Month6

Pickmodels. Build features. Looks great. Demo.toleaders. Sign-off. Ship. b's'-ing us?" "WhyisAl failedlastyear. sSin projects

Sound familiar? You're not here because you haven't seen this. You're here because you want to stop it.

AlEngin

CURCPE

AEngivoer Engineering thefuture of Al

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: agent_vision.

Slide text:

The AI is the easy part

You can't debug what you can't see.

You can't improve what you can't measure.

You can't trust what you can't explain.

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: agent_vision.

Slide text:

The Five Pillars of Production AI

Evaluation

Observability

Data Foundation

Orchestration

Governance

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: agent_vision.

Slide text:

Evaluation First

Define success with numbers

Build test cases from real data

Wire automated grading

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: advanced OCR rapidocr-live/bright-screen/contrast.
OCR decision: ready — multi-column dense text and embedded code screenshot are better handled by OCR

Slide text:

PILLARO1-DEEPDIVE

Three layers of evaluation

Layer1—Deterministic

AIE PlI detection (NER+regex),Output format validation,Response length bounds aremalgretaltAi（Mlwtrepote

wrng.+gatiallyorret,2·foil（rect

Layer 2 -- Semantic Correctness & groundedness,LlM-as-a.Judge,Non- above threshold determinismfix:runeachtest3x-flagvariance utety: wtr:n. *11: 1evtyktelclasooreoytxretlenteeet！ Dees tte resposie snsld Plt leakage asd ballocinated sccoor dsta)

C.5 tevest scefe.

AlEngin escalate when confidencewas low?Did it stay Layer 3 - Behavioural Did it call theright toois,in the right order?Did it within scope? 5D.*** (entamtgoeni（oary) RtrN（ometr(（anteet) A1re1ponse:(re1gcsse） SAMPLE: LLM-aS-a-Judge prompt

RURNPE

ABrgore AI Engineer

EUROPE

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: advanced OCR rapidocr-live/center-82/contrast.
OCR decision: ready — Dense architecture diagram with many small labels and stacked elements.

Slide text:

Databricks Data Intelligence Platforn

Dissster recovery 100%serverless Cost controls Enterprise security

AIE Artinicial inteigence Mosaic Al Databricks SQL Data warchoushg Workflows/SDP IngesETL streamng e Business intelgonce AI/81

Lakehouse

Unity Catalog

AlEngin DELTA LAKE ICEBERG Parquet

AEngineer AI Engineer

2028 EUROPE

Recreated text/layout view: open HTML recreation
AI slide classifier: demo_video confidence 0.88
Text source: agent_vision.

Slide text:

Engineering the future of AI

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.99
Text source: advanced OCR rapidocr-live/border-trim/opencv-adaptive.
OCR decision: ready — Dense platform architecture slide with many small labels and nested sections.

Slide text:

databricks

Agent Breks rteilgerce A1/BI Agentrc business Secue data snd Al sppy Custom Apps And more.

★ + ★ AIE? Reasoning Agents Contextual Developer Platform Agent Platform Al Governance

Knowicdge Assistant Agcnt Orchostration Agcnt/Skill/McP Rogistry

Supervisor Agont Runtimo AlGatcway

Documonts Agent Memory Agont Observability

AlFunctions Copacity Modol 0 Gemn Al Managed OAuth Apps

FAEngk

AEngrak AI Engineer

EUROPE

Recreated text/layout view: open HTML recreation
AI slide classifier: content_slide confidence 0.98
Text source: agent_vision.

Slide text:

Same project. Same problem. Different approach.

18,000 calls/month - 60% simple queries

$85,000 + 6 months spent on failed attempt

System: unmeasurable, invisible, misaligned

Goal: AI agent handles 60%+ user queries

We didn't pick a model until week 7.

Hidden Non-Slide Evidence

`slide-001.jpg` — speaker_stage confidence 0.99; camera shot of presenter on stage with audience and only a partial projected slide visible
`slide-011.jpg` — speaker_stage confidence 0.99; Camera shot of the speaker on stage with a projected slide in the background, not a readable presentation slide.

Classification audit: raw/sources/slide-ai-classification/slides/ObTPqBGsEbA/audit.json

Slide-Derived Subjects To Review

Subject extraction uses video title, related session titles/descriptions, transcript context, and OCR text when available. OCR is best-effort and should be reviewed against the embedded slide images.