Slides: Make your LLM app a Domain Expert: How to Build an Expert System — Christopher Lovejoy, Anterior
Source Video
Make your LLM app a Domain Expert: How to Build an Expert System — Christopher Lovejoy, Anterior
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
OCR text:
INNOVATIONPARTNER
aws
PLATINUMSPONSORS
Graphite
WWindsurf
MongoDB
daily
augment code
Workos
OCR text:
Own yourvertical Aolaysook for buildinga o
comain-native LLM application ha,
i”
i oO
| a Microsoft Qo)
OCR text:
Anterior
We're a New York-based, clinician-led
company that provides clinical
reasoning tools and solutions to
accelerate and automate healthcare
administration
OCR text:
OUP BET FOR VERTICAL AT APPLICATIONS:
an
LE ee ese
5 ra
SYSTEM Tor an pane Lae Megas yaseaeaacdllet
; le : SOONISTICATION
I RICS Se Olee kel ale we ay veliaane aia
ate tetclealaicie igi een eee
ec and pipelines
c
OCR text:
Wirvy is tt arc! to successfully agply
| Lis to specialized incustries’
mvt i)
| mw Microsoft §GDOU
OCR text:
An exampic cinical case processed by
On eee ne oa ee eet
Gh ls there documentation of unsuccesstul
merinpakvic tarrearis fae atk aan ae
conservative therapy for at least 6 weeks?
E -—T, i)
u Microsoft §=Gocl)
OCR text:
We achieved a paseline performance of 95% tor approvind Care...
| V1 fai)
mw Microsoft § SOU
OCR text:
KLAS CASE STUDY
We achieved a baseline performance of 95% for approving care...
then iterated with a specific customer to >99% within 8 weeks
Initial Results 8 Weeks Later
Performance (F1-Score)
95.73% Performance (F1-Score)
99.24%
using the system
outlined in this talk
OCR text:
(2) Empower domain experts to define and
maintain a failure mode ontology
Photo extraction
Table extraction
Handwriting extraction
Medical record extraction
Checkbox extraction
Medical Necessity Review
Failure Modes
Logic representation
Rules interpretation
Rule source selection
Clinical reasoning
Under-inference
Over-inference
Chronological reasoning
OCR text:
(3) Measure eras together
on oroduction data to uniock tarceted Insiqnts
; i. a . 7 ; oe a os
Ree eee ae
i | 7 |
| ' wick aS 000 Oh
OCR text:
AIE
MEASURE
SpecfyFailuremode'tofx
Makechanges
ReviewAIoutputs
Analyze perfornance
toprioritisework
Produetion
Performance
opplication
insights
"Failure mode
dataset"eval
Chnica!
AI
ChnicalPA
engineer(s)engineer(s)
Make decisions
aboutnaking ive
See inpact of changes
aws
OCR text:
Failure mode datasets enable targeted product
iteration
OCR text:
AIE
Empower your domain experts to make
improvements directly with tooling and evals
Application Pipelines
Tooling for
domain experts
Domain evals
Domain
Knowledge Base
Microsoft
smol.ai
OCR text:
AIE
Putting it together
Microsoft
smol ai
OCR text:
Takeaways
• To build domain-native LLM applications you need to solve the last mile problem.
• Using the best models isn't enough - you should build an adaptive domain intelligence engine.
• Domain experts power this system by reviewing AI outputs to generate performance metrics, failure modes and suggested improvements.
• This takes production data and uses it to give your LLM product a nuanced understanding of customer workflows.
• The result is a self-improving, data-driven process that can be managed by a domain expert PM.
OCR text:
AIE
Thankyou
pherovejyMD
chrislovejoy.me
Anterior.com/Company
Microsoft
smol?
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.