Leo Teixeira
LinkedInGitHub
PTLet’s talk

AI, product, and engineering cases in real-world context

Complexity turned into something people can use.

A layered reading: cards for orientation, quick views for context, and full cases for a deeper look.

08 visible works
01 / AI case · Internal productInternal tool in use
Visual conceitual do workspace PromptHub, com editor de prompts, respostas comparadas e trilha de avaliação.

PromptHub

Internal platform that combines editing, multi-LLM testing, and Git versioning to turn scattered prompts into traceable, comparable, governed assets.

roleArchitecture and implementationfocusPrompt governance and evolution
Prompt EngineeringMulti-LLMGovernance
02 / AI case · Applied architectureAnonymized solution in a real-world context
Visual conceitual do Assistente RAG multiproduto, mostrando uma pergunta roteada entre bases de conhecimento até o atendimento humano.

Multi-product RAG assistant

RAG-based assistant able to serve different products from specific knowledge bases, preserve context, and hand off the conversation when automation is not enough.

roleTechnical leadership and architecturefocusContext and human continuity
RAGRoutingHuman support
03 / Applied research · CESAR SchoolMaster's research · CESAR School
Visual conceitual da pesquisa em Engenharia de Prompt, com comparações experimentais e sinais de confiabilidade.

Engenharia de Prompt, Evals e Confiabilidade

Applied research comparing Prompt Engineering techniques to improve the quality, reliability, and efficiency of solutions based on language models.

roleSoftware Engineering researchfocusEvaluating LLM-based solutions
Applied researchEvalsLLMs
04 / AI-first product · Product EngineeringMVP in progress
Visual conceitual do SpecFlow, transformando notas de discovery em uma especificação estruturada e rastreável.

SpecFlow

AI product studio that turns raw discovery notes into structured specifications, a prioritized backlog, and a risk matrix with literal traceability.

roleProduct, architecture, and implementationfocusTraceable specification
Applied AIDiscoveryZodEvals
05 / Own product · AI-firstOwn product
Visual conceitual do laboratório de produto e IA Spyke, com painéis conectados e um núcleo de capacidades.

Spyke

AI-first product conceived as a lab for integrating Artificial Intelligence, product, architecture, and experience from problem definition through implementation.

roleProduct, UX, and architecturefocusProduct and AI lab
AI-firstOwn productArchitecture and UX
06 / AI experiment · Applied RAGPublic demonstration
Visual conceitual do AskYou, conectando pergunta, contexto recuperado e resposta fundamentada.

AskYou

Demonstration chatbot that combines embeddings, vector search with FAISS, and Google Gemini to answer questions from a synthetic knowledge base.

roleArchitecture and implementationfocusAnswers with retrieved context
RAGEmbeddingsFAISSGemini
07 / Demonstration product · Applied AIPortfolio prototype
Visual conceitual do TalentScore AI, organizando currículo, sinais de análise e próximos passos.

TalentScore AI

Product prototype that organizes resume analysis into scores, strengths, gaps, keywords, and actionable suggestions for people seeking opportunities.

roleProduct, UX, and architecturefocusActionable resume feedback
Next.jsAI productDashboardFeedback
08 / Experimental product · Machine LearningExperimental prototype
InnPricer

InnPricer

Experimental project about hotel pricing with Machine Learning and Artificial Intelligence, connecting data, product hypotheses, and an experience for investigating price.

roleExploration and prototypingfocusExplore hotel pricing with ML and AI
Machine LearningPricingHospitalityAI

PLACEHOLDERSome visual details and evidence will be added when there is context and permission to publish them. Sensitive claims follow the same principle.