Turning Regulate Data into an AI Ingredient Intepreter
Overview:
I designed a skincare ingredient interpreter built behave like a data pipeline. The goal was to create a tool that explains regulated ingredients plainly, stays grounded in verifiable data, and never fills gaps with guesswork. The system has controlled inputs, consistent processing, and transparent outputs that are user friendly.
The Problem:
Ingredient information is scattered across different regulatory bodies. The bodies don’t always align on naming, formatting, or recommended usage. Most consumer-facing tools resolve these inconsistencies by smoothing over the gaps or use simple language that reflects marketing terms over actual data.
If something is unknown, the consumer tool should acknowledge it. If information conflicts, the system should show the conflict, not bury it. My challenge was to build an ingredient interpreter that stays within the boundaries of science and doesn’t make up information.
Research & Insight:
To understand the landscape, I gathered ingredient data from reliable scientific and regulatory sources and standardized it. This process exposed how inconsistent the raw information was, and a lot of back end logic and protocols were was need to keep the data consistent.
The clarity of the explanation is directly tied to the clarity of the underlying structure. When the data model is consistent, the experience becomes consistent.
System Design:
I created a single schema that defines every ingredient field the system is allowed to use. This schema acts as the source of truth for the entire platform. When it updates, everything else updates with it: the database, the endpoints, and the interpretation format.
This prevents drift, reduces ambiguity, and keeps the system aligned over time.
Consistent Processing:
Each ingredient moves through the same pipeline:
Bring in the raw data
Clean and standardize it
Check it for contradictions
Review and validate
Lock it with a version identifier
If anything changes, the version changes. This makes every explanation traceable.
Structure-Generated Endpoints:
The platform’s endpoints are generated from the schema. Nothing is hand-written. This ensures that what the system stores, serves, and presents is always synchronized.
The interpreter uses the structured ingredient record as its sole reference. If information conflicts, the explanation shows the conflict so the user can understand the boundaries of what’s documented. The tone remains steady and factual, without alarms or false certainty.
Testing:
I ran 50 ingredient-related questions to pressure test the system: safety, compatibility, concentration limits, edge cases, and messy queries real people might ask.
Results:
Zero hallucinations
High factual consistency
Clear fallbacks when information was incomplete
Deviations tied to missing fields, not model failure
Outcomes:
Most skincare tools rely on loose databases and optimistic phrasing. They can’t explain how they reached an answer. My final system behaves more like a regulated information service than a beauty tool. Every response ties back to a versioned dataset and a defined structure.
Stable, verifiable ingredient information
Consistent explanations tied to real data
Clear boundaries between what is known and not known
Full traceability through versioning
Next Steps:
Automate regulated data gathering
Expand coverage to 80+ ingredients
Test with non-expert users
Refine explanation depth and tone
The goal is to keep the interpreter both rigorous and approachable as it expands.
Reflection:
This project is an example of how I design high-trust systems: start with structure, not gloss. Give the AI guardrails rooted in data, not vibes. Build something that stays honest even when the answers aren’t flattering or simple.
Good UX in regulated spaces isn’t loud. It’s steady. It’s traceable. It respects the user enough to tell the truth.