Question
How can teams make AI portfolio and inference-cost decisions with a clearer view of controls, assumptions, and usage?
Context
Why this mattered
A Streamlit app bringing portfolio framing, control concepts, and inference-cost reasoning into one decision-support workspace. Modeled estimates are not presented as production benchmarks.
Approach
From ambiguity to evidence
Defined a portfolio view linking use cases, model choices, usage drivers, and controls.
Added inference-cost exploration around transparent inputs and assumptions.
Designed an interactive flow for comparing scenarios and documenting decision context.
Findings
What the analysis surfaced
Cost conversations are clearer when usage assumptions and model choices are visible together.
A control-oriented view helps separate governance questions from optimization questions.
Outputs depend on user-provided assumptions and are not production cost benchmarks.
Contribution
What I owned
Product strategist and analytics modeler; defined the decision flow, scenarios, and interface requirements.