SuperCharge / AI development
Building with AI.
I co-created and led SuperCharge, bringing AI-assisted development into real team workflows. I drove architecture, infrastructure, implementation, and the roadmap, working with engineering teams and leadership on adoption.
SuperCharge grew into a major Dot Foods IT initiative. My work also included introducing practical workflows to colleagues and preparing the roadmap and handoff to the team carrying it forward.
- Idea
- →Implementation
- →Team adoption
AI search / Ecommerce
Smarter Search.
I originated and architected the reusable AI search-recovery foundation behind Product Matching. That foundation supports both GTIN and Dot Item matching, with implementation by a colleague.
GTIN matching helps customers find alternative products when they search for a GTIN that Dot does not carry. Dot Item matching applies the same foundation to inaccessible or long-discontinued Dot items.
In its first 21 weeks, GTIN Product Matching was used 18,816 times by 2,139 external users across 935 accounts.
For external visitors searching 11-, 12-, and 14-digit product codes, the unresolved no-result rate was 11.6% lower than in the preceding 21 weeks.
11.6% lower no-result rate. Numeric product-code searches · 21-week before/after comparison.
One foundation. Two matching tools.
- GTIN matching
- $349,290modeled annual net margin
- Dot Item matching
- $41,786modeled annual net margin
Fuzzy search
I also implemented fuzzy search to handle misspellings and search-term variations. Reported year-to-date no-result searches declined 29.3% year over year following rollout.
- Fuzzy search estimate
- $53,763modeled annual net margin
Conditional model estimates, not verified additional profit. Feature effects may overlap; the amounts are not additive.
- Unsuccessful search
- →AI matching
- →Relevant products
