CTO at Applause.team · Co-founder at MokLabs
I build technical organizations and the systems that let them move.
matheuskindrazki.dev ·
LinkedIn ·
Medium ·
X ·
Email
Curitiba, Brazil · −25.42° −49.27°
I'm a CTO, entrepreneur, and engineer who spent nearly a decade architecting platforms used by millions of students — and is now turning that experience into high-agency teams and products of my own.
I think in systems. I like understanding a problem deeply before touching it, then building the kind of solution that holds up: predictable, observable, documented, and automated so the right thing becomes the easy thing. I care about results that are real and durable over things that just look impressive — mechanisms over magic.
Off the clock I'm a musician and a DJ, the piano is my favorite instrument, and — turns out — I also know how to make rubber. Creativity doesn't stop at the code.
Applause.team — CTO. A gamified sales incentive and recognition platform. I lead technology across product architecture, engineering systems, team velocity, reliability, and the operating model that lets the organization scale.
MokLabs Venture Studio — Co-founder. An AI venture studio that turns sharp ideas into shipped products: shared platform engineering, design systems, and AI infrastructure so each new venture reuses the scaffolding instead of rebuilding it.
Lugui.ai — Founding team. AI infrastructure for the real-estate market — "the intelligence of your agency, inside WhatsApp." Conversational agents and RAG pipelines that let realtors qualify leads, answer about listings, and close deals where their clients already are.
CEIA (UFG) — AI Research Fellow. Applied research at the frontier of LLMs and intelligent systems, bridging academic rigor and shipped products.
Jarvis — My personal knowledge graph + AI agent. A private second brain on MCP servers and vector search, connecting memory, entities, tasks, and decisions across every project I touch.
- FastShared — Ephemeral file sharing as a single gesture: pick a file, set how long the link lives, get a URL that vanishes on its own. Apple-native + a CLI built as a file-handoff primitive for AI agents.
- Lavra — A native macOS app that transcribes audio to clean text fully offline on Apple Silicon (WhisperKit + Apple Foundation Models). Strong Portuguese-BR support; nothing leaves the device.
- ScreenshotDetox — Turns your screenshot pile into an inbox of decisions, entirely on-device (Apple Vision OCR + a local classifier).
- Synk — A real code editor for the iPad with on-device AI (SwiftUI + Runestone, cloud AI plus an opt-in local model via MLC-LLM).
- Remindr.AI — Privacy-first desktop app for intelligent transcription, diarization, and daily memory — for people who think while they talk.
- Kindraw — A warm, playful workspace on top of Excalidraw: authenticated boards, markdown docs with Mermaid, and real-time collaboration.
- Lofiever — A 24/7 lo-fi streaming app with an AI DJ that curates the playlist and talks back in chat.
Nearly six years at one of Brazil's largest edtechs. I joined as a Software Engineer, stepped up to Staff with microfrontend architecture, and consolidated as Principal Engineer of AI.
- Architected the microfrontend ecosystem with Module Federation, enabling independent squad velocity while keeping system coherence across product surfaces reaching millions of students.
- Led design-system integration and governance, plus DX — canonical documentation, RFC/ADR processes, CI/CD standardization.
- Last delivery: the AI essay-correction platform for the ENEM exam. I led the core pipeline (ingestion, classification, analysis) and owned the computer-vision model that reads handwritten essays (OCR/HTR) over a queued, real-time correction system feeding teachers.
- Seeded Café com Código, a recurring engineering ritual for knowledge sharing and technical culture.
- Coherence over cleverness — systems should be predictable, not impressive.
- Governance as enabler — lightweight structures (RFCs, ADRs) that prevent chaos without bureaucracy.
- DX as leverage — remove friction and teams move faster with higher quality.
- AI as engineering — observability, evaluation, datasets, and metrics; not magic.
- Documentation as source of truth — canonical, versioned, discoverable.
- Pragmatic depth — deep enough to be correct, simple enough to be maintainable.
Frontend Architecture & Platform — React · Next.js · TypeScript · Module Federation · Micro Frontends · Design Systems Applied AI & Agents — LangChain · LangGraph · RAG Orchestration · MCP Servers · Vector DBs · On-device AI · OCR/HTR · Computer Vision Full-Stack & Native — Node.js · Python · PHP · Rust · SwiftUI · Tauri · GraphQL · PostgreSQL Cloud & Infrastructure — AWS · Azure · Cloudflare · Docker · Coolify · Supabase · Edge Functions
Building coherent systems, one mechanism at a time.