← Back to projects

AI / Retrieval System Design

Chemical AI

Designed an SDS-focused AI assistant for chemical safety workflows using structured document ingestion, metadata extraction, retrieval pipelines, and safe-response design.

RoleSystem designer / AI engineer
Timeline2026
PythonFastAPIPostgreSQLRAG

Overview

Chemical AI is an applied AI project centered on helping users retrieve and understand chemical safety information through SDS documents and structured retrieval.

The challenge

The challenge was to design something genuinely useful and safety-aware, not just another chatbot layered over generic documents.

What I built

Planned structured SDS ingestion and metadata extraction workflows.
Designed retrieval pipelines for surfacing relevant information from safety documents.
Considered safety guardrails, response constraints, and escalation behavior.
Built the project around realistic data architecture and domain-specific information needs.

Outcomes

Demonstrated applied AI thinking beyond simple prompt engineering.
Showed understanding of retrieval architecture, structured data, and safety-aware system design.
Created a strong foundation for a future production-oriented chemical safety assistant.