Prize Categories Entered

• 

  Best Use of Gemini API

  Submitted

  Gemini powers Skimple’s multimodal product scanning. Through the Google Gen AI SDK, we send front packaging photos to Gemini and request structured JSON containing the product identity and ordered ingredient list. The extraction flags illegible fields and incomplete ingredient lists instead of filling gaps with guesses. Users confirm the result before analysis. Gemini also classifies firsthand experiences from permitted review sources into structured aspects and sentiments. This turns packaging photos and unstructured review text into validated inputs for our product reports.

• 

  Best Use of Tiger Data

  Submitted

  Tiger Data’s PostgreSQL database serves as Skimple’s system of record, connecting skin profiles, product formulations, ordered ingredients, scientific papers, reviewed claims, review observations, scans, and saved reports. We combine relational tables with JSONB for flexible questionnaire answers, model extractions, and report data. Versioned formulations preserve which ingredient list each analysis used, while cached review summaries reduce repeated external API calls. This gives Skimple one database for personalized product analysis, source provenance, and persistent user shelves.

• Event prize

  Best Overall

  Submitted

  Skimple helps people understand weather a product suits them and what's best for their skin without spending hundreds of dollars. User photographs the packaging and Skimple confirm the ingredients, and get a report based on your skin goals, sensitivities, and known reactions. Gemini reads the label. Nimble finds current product information and relevant reviews. TigerData PostgreSQL connects formulations, research, and your profile. Sentry helps us find failures across the process. Skimple shows the sources behind a finding, flags ingredients you’ve reacted to, and says when the evidence is thin. That gives people a reason to trust the answer and a way to question it.

• Event prize

  Best Use of Nimble

  Submitted

  Nimble connects Skimple’s product scans to live web context. After reading the Product name, we use Nimble’s Search API to find ingredient listings for the identified brand, product, and variant. We also use Nimble to get relevant product reviews from across the web. We also use nimble to get the ingredient list of a product as well as get research data on the ingredients, based on actual research papers.

• Event prize

  Sentry Raffle

  Submitted

  For Skimple we do image extraction, web retrieval, database matching, and background analysis. We implemented our db with Sentry transactions to cover those stages. It helps us locate failures and debug steps across Gemini extraction, Nimble searches, formulation matching, review classification, and compatibility scoring. We configure Sentry with default PII collection disabled and request-body capture turned off, keeping sensitive inputs out of monitoring while retaining useful diagnostic context.