Skip to content
View leo-statai's full-sized avatar
  • Campinas, São Paulo
  • 23:13 (UTC -03:00)

Highlights

  • Pro

Block or report leo-statai

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
leo-statai/README.md

Leonardo Alves

Building AI systems with the rigour of a Data Scientist and Statistician.

Based in Campinas (SP), Brazil. Professional experience as Data Scientist and Statistician across consulting, market research, scientific research, and the consumer-goods industry. Now focused on AI engineering — building LLM-powered systems end-to-end, with attention to evaluation, governance, and reproducibility.


Featured work

Project What it does
ai-generated-content-evaluator LLM pipeline that generates technical reports with NotebookLM and auto-evaluates them with the Gemini API across four quality metrics.
apertus-ethics-by-design-case-study Case study mapping the Swiss Apertus LLM to the EU Ethics by Design framework and AI Act, with a quantified compliance/performance trade-off analysis.
whisper-transcriber Self-hosted web app for audio/video transcription on NVIDIA GPUs — SvelteKit + FastAPI + Redis/ARQ + faster-whisper, resumable uploads via tus, real-time progress via SSE, 5 export formats.
ai-fluency-ptbr Brazilian-Portuguese translation of A Framework for AI Fluency (Dakan & Feller), with an interactive SPA companion (Tailwind + Chart.js). Live demo.

Toolkit

  • LLM APIs: OpenAI · Anthropic · Google Gemini · DeepSeek
  • Local LLM stacks: Ollama · AnythingLLM · Open WebUI
  • AI agents: Hermes Agent (building custom agent profiles on a local server)
  • AI-assisted development: Claude Code · Codex (both CLI)
  • Languages: Python · R · SQL · LaTeX
  • Web & backend: FastAPI · SvelteKit · Docker
  • Speech & audio: Whisper · faster-whisper
  • Linux & home lab: Debian · Ubuntu · self-hosted services

Education

  • 🎓 B.Sc. in Information Technology — UNIVESP, graduating June 2027
  • 🔬 PhD-level coursework (special student) at UNICAMP / FEEC: Responsible & Ethical AI (IA364) · Seminars in Computer Engineering (IA382)
  • 🎓 M.Sc. in Statistical Modeling — UNICAMP / FEA Multivariate regression with Partial Least Squares (PLSR), applied to Sensory & Consumer Science.
  • 🎓 B.Sc. in Statistics — UNICAMP

Certifications

  • Machine Learning Specialization — DeepLearning.AI / Stanford Online (Coursera, 2025) Supervised ML · Advanced Learning Algorithms · Unsupervised Learning, Recommenders & RL
  • 5-Day Gen AI Intensive Course — Google × Kaggle (2025)
  • Google Data Analytics Professional Certificate — Google / Coursera (2024) SQL · Tableau · R · spreadsheets
  • Statistical Learning, with Distinction — Stanford Online (Hastie & Tibshirani, 2020)
  • Data Science Specialization — Johns Hopkins University (Coursera, 2017)

Get in touch

📫 [email protected] — open to roles and collaborations in AI Engineering, Applied AI, AI Data Science, AI Data Analytics, and AI Governance.

Pinned Loading

  1. ai-fluency-ptbr ai-fluency-ptbr Public

    Brazilian-Portuguese translation of A Framework for AI Fluency (Dakan & Feller), with an interactive single-page web companion (Tailwind CSS + Chart.js). Live demo: https://leo-statai.github.io/ai-…

    HTML

  2. ai-generated-content-evaluator ai-generated-content-evaluator Public

    LLM pipeline that generates technical reports with NotebookLM and auto-evaluates them with the Gemini API across four quality metrics.

    TeX

  3. apertus-ethics-by-design-case-study apertus-ethics-by-design-case-study Public

    Case study mapping the Swiss Apertus LLM to the EU Ethics by Design framework and AI Act, with a quantified compliance/performance trade-off analysis.

    TeX

  4. focus-study-tracker focus-study-tracker Public

    Self-hosted study time tracker — Python stdlib + SQLite, browser-based, LAN-accessible, Docker-ready.

    Python

  5. whisper-transcriber whisper-transcriber Public

    Self-hosted web app for audio/video transcription on NVIDIA GPUs — SvelteKit + FastAPI + faster-whisper, resumable uploads, real-time progress, multi-format export.

    Svelte

  6. transcritor transcritor Public

    Script em python para transcrever arquivos de áudio e vídeo para texto.

    Python 4