Hackathon Portal
AI Tinkerers - São Paulo
Team

ALRY

Project Concept

Context Problem: in big hospitals, more than 10,000 tests per day are done; it means that the nurse or the doctor takes approximately 10min per test, or 100.000 per day/person.

Current State: Visual interpretation of text (taking a nurse approximately 10 minutes per analysis, without AI assistance). The patient could have access to a tailored interpretation of the exams in non-medical terms.

Objective: To assist physicians and patients in interpreting medical exam results using AI and UI Generative.

Use Cases:

Single Blood Test Interpretation: Assist with basic interpretation.

Multiple Blood Test Interpretation: Assist with interpretation and identify patterns in blood levels.

Knowledge Base Integration: Assist with interpretation, identify patterns, and leverage a scientific knowledge base to highlight potential health risks for the physician, backed by medical literature.

Example: With 5 consecutive blood tests, the AI could detect a trend and suggest potential conditions to the physician.

Solution: A plataform where the user input a pdf (e.g. hemogram result) and get the analyzed result of the ina generative UI interface. You can input many hemogram result to the AI recognize pattern and point out disease risks and attention. It also helps physicions to interpret the results in a user friendly interface and the patient can also have a non-medical interpretation of this result.

Provide seamless access to interpretations, consultations, and detailed exam insights. With the use of AI and

Possible Impacts:

Time Saving: Generative Interfaces (Generative UI), the interpretation time for a normal complete blood count (CBC) can drop from 10 minutes to less than 30 seconds.

Quality of the interpreation

User friendly interface of the blood results

Benefits: medical interpretations of the exams are more optimized and interactive; more user-friendly interface for nurses and doctors.

Escalability: this solution can be applied in human medical and veterinary medical context.

Entry

Status: Submitted

Last saved: May 16 at 4:59 PM -03

Team Roster

You must be registered for the event to view the team message board.

Yuki Tanaka Team Lead RSVP Approved

PO at Stefanini
Product vision and strategy.
Yuki Tanaka is a hackathon lover: have participated in 16 hackthons already since the end of 2024. Product Manager and Product Owner based in Curitiba, Paraná, Brazil, currently driving product strategy at Stefanini Group. With six years of experience, Yuki blends technical expertise in analysis and systems development with a solid grounding in financial accounting and even veterinary medicine, earned through studies at Senac, Universidade Estadual Paulista (UNESP) and the University of São Paulo. At Stefanini, Yuki leads cross‑functional teams to deliver innovative solutions, translating market needs into actionable roadmaps. Passionate about bridging technology and business, Yuki continuously seeks to enhance product performance and user satisfaction.
LLM/chatbot design (DialogFlow CX, Vertex AI), generative-AI chatbots, LLM-based platform capabilities, R&D product pipeline & backlog governance, CRM integrations (Salesforce, Medallia), omnichannel CX & NPS, data-driven product decisions (SQL/analytics), prototyping/UX, API integrations, Agile product discovery, roadmaps & OKR definition
Projects: R&D product-pipeline coordination — managing intake, prioritization and release planning; backlog governance — sprint readiness, acceptance criteria and QA; LLM platform enablement — piloting Dialogflow CX and Vertex AI chatbots, reusable prompt templates and Vertex deployments; CRM integrations — building Salesforce & Medallia connectors and SQL analytics for NPS/omnichannel CX.

Ana Lindiner Lima Araujo RSVP Approved

Data Engineer at BIP Brasil
Development using Copilot Kit.
Sou engenheira de dados com experiência em projetos que unem tecnologia e sustentabilidade. Atuo em iniciativas voltadas para a redução de emissões de carbono no agronegócio e na construção civil, desenvolvendo soluções digitais inovadoras que promovem impacto positivo no meio ambiente. Sendo que trabalho com equipes e projetos estratégicos, sempre com foco em inovação, transformação cultural e responsabilidade climática. Minha trajetória inclui a co-fundação de empresa júnior, participação em programas de liderança e atuação em consultoria estratégica. Meu propósito é inspirar e transformar por meio da tecnologia, preparando organizações e pessoas para o futuro sustentável impulsionado pela IA.
Data engineering, sustainability, carbon emissions reduction, agribusiness, civil construction, digital solutions for sustainability, innovation, cultural transformation, climate responsibility, strategic consulting, AI-driven sustainable future
Projects: carbon-emissions reduction platform for agribusiness — building data pipelines, ETL and analytics to measure and report emissions; carbon-tracking solution for construction — integrating project data and producing sustainability dashboards and KPIs; internal sustainability data platform to support strategic consulting engagements and client reporting.

Ricael Menezes Durand RSVP Approved

Coordenador de Backend at Till Tecnologia e Mobilidade
Development using Copilot Kit.
Ricael Menezes Durand is a Software Developer , currently serving as Coordenador de Backend at Till Tecnologia e Mobilidade in São Paulo, Brazil. He is employed and open to new opportunities. Ricael's expertise lies in coordinating backend architecture for high-volume, real-time systems, with a focus on Node.js, Express, and WebSockets. His technical projects include building and maintaining microservices using Clean Architecture and Domain-Driven Design. Previously, he developed projects like MovieFlix and library management systems with Java (Spring Boot), Hibernate/JPA, and PostgreSQL, utilizing Docker and GitHub Actions for deployment and CI/CD. He studied Programação de Computadores at Instituto Federal de Educação, Ciência e Tecnologia de São Paulo - IFSP. Ricael is open to introductions via direct message.
Backend architecture, high-volume real-time systems, Node.js, WebSockets, microservices, Clean Architecture, Domain-Driven Design (DDD), Java Spring Boot, CI/CD, engineering quality, mobility marketplaces.
Currently coordinating the backend architecture for a high-volume, real-time mobility marketplace at Till, focusing on Node.js, Express, and WebSockets. Technical projects include building and maintaining microservices using Clean Architecture and Domain-Driven Design (DDD). On GitHub, active work includes MovieFlix and library management systems developed with Java (Spring Boot), Hibernate/JPA, and PostgreSQL, featuring containerized deployments via Docker and CI/CD through GitHub Actions.

Lucas Martins de Oliveira RSVP Approved

AI Engineer at APEX Partners
Development using Copilot Kit.
I'm Lucas Oliveira, a 21-year-old AI Engineer from Vitória, Brazil. I started studying AI at 17, while living in Boston, when I was invited to take an introductory course on Python, AI and Data Science. I fell in love with the field and have been learning ever since, focused on building solutions that help society across multiple domains. I currently lead ARI, a large-scale economic intelligence platform at a Brazilian investment bank, and I'm pursuing a B.A. in Economics at FUCAPE. My work spans medical, financial, operational and empirical solutions. I draw deep inspiration from my family, including a relative who teaches machine learning at ESPM. Outside tech, I played competitive tennis for years, which shaped my discipline and exposed me to diverse cultures.
I'm drawn to the intersection of AI, finance and entrepreneurship, specifically the unsexy gaps incumbents overlook. Brazilian and LatAm markets are full of blind spots: lay users locked out of digital products, small firms priced out of Bloomberg-grade intelligence, regional economies underserved by data. I want to go deeper into LangGraph agents and LLM token economics, and connect with founders building vertical AI for finance and operators tackling underserved markets.
I'm building Simpl Ai, a conversational website builder targeting non-technical Brazilian users, integrating backend, payment gateways, third-party tools and domains end to end. In parallel, I lead ARI, an economic intelligence platform trained on 2M+ records from consolidated Brazilian sources. Stack: LangGraph, AWS, OCI, Anthropic SDK, Codex, Docker, Railway, Langfuse, pgvector, FastAPI, Hugging Face.