Intimus “She Can” NFT art festival (Latin America)
Idea co-owner, solution architect and full-stack engineer of the first marketing campaign utilizing NFTs for community building as part of Intimus’ 3 years strategy to build digital communities.
Solution: High-fidelity prototype for consumer validation using a Next JS front-end with Flowbite components (Tailwind) and EVM Smart Contract managed through ThirdWeb SDK.
Tech Stack: ThirdWeb / Wagmi / Connect Kit / Alchemy / IPFS / NextJS / Azure.
“Proof of Achievement” Non-fungible Awards (Global)
Idea owner, solution architect, solidity engineer and product owner of an internal experiment to “de-risk” and show value of Web 3 technologies to businesses and Global IT leadership. Award NFTs in 2022: “CIO Awards”, “R&E Technical Excellence” and “Global Digital Innovator of the Year” OpenSea.
Solution: EVM Smart Contract Factory deployed on Polygon (L2) to manage Smart Contracts (awards) with ERC1155 tokens (multiple owners).
Tech stack: Hardhat / Alchemy / Piñata / Wallet Connect / React / Storybook / Web3.js / GraphQL / Azure.
“Point of View” (POV) on Metaverse (Global)
I discuss the potential of decentralized metaverse and blockchain networks, as well as the possibilities created by decentralized applications and non-fungible tokens. The POV article explores how these emerging technologies can be used to innovate and create value for CPG corporations. By leveraging the reduced cost of verification and networking of non-fungible tokens and decentralized applications, CPG corporations can create new revenue sources and better engage with consumers.
Solution: A point of view article presented to the CIO organization and the Global Digital Technologies Solutions organization on the evolution of the concept Metaverse, diving into some of its building blocks and how the maturity of decentralized blockchain networks have contributed to a alternative proposal of what the Metaverse will be. It also looks at the functional role of crypto and non/semi-fungible tokens.
Computer Vision Model and APIs (Global)
Solution architect and ML engineer to train a transformers model for classification of poop color, served through APIs on Azure to a React front-end pilot application poposcan.com
Solution: This initiative has 3 distinct architectures that have been developed to test different hypothesis and provide options for markets with specific needs. First version was developed in-house using a React Front-end, Azure serverless functions behind API management and a non-relational DB. Second version was built by a partner with computer vision expertise, which used images from our R&E to train a classification model and also provisioned a newer version of the front-end. Third version mocks up the machine learning model using RGB color maping to cut costs and reduce maintenance needs, priding very similar results. A in-house transformers model has also been trained using data collected from real consumers from the previous versions with very positive results.