Role and product problem
Vladimir Fedorov co-founded ShelfSignal, designed its architecture and built the core pipeline. The product measures how assistants name brands and which sellers they select in simulated shopping briefs.
Engineering approach
The visibility workflow collects repeated answers to buyer questions and extracts brands, order and cited sources into structured records. A rules engine links observations to recommendations. A separate shopping simulation records constrained brand and seller choices and stops before payment.
Documented scope
The published category format uses 20 questions, five assistants and five repetitions:500 answers. The portfolio reports 228 brands surfaced in one magnesium run and about$20 in API costs for that run. Those are sample observations, not guaranteed costs or proof of a customer’s sales lift.
Client accounts, self-serve reports and dashboards complete the application around the measurement pipeline.
