I build production-ready AI systems that connect machine learning, optimization, data engineering, and full-stack development.
I'm an AI/ML Engineer and Data Scientist from Manaus, Brazil, focused on building end-to-end intelligent systems. My work ranges from experimentation and model development to scalable APIs, data pipelines, cloud deployment, and user-facing applications.
- 🧠 Building AI agents, RAG workflows, predictive models, and optimization systems
- 🏭 Applying AI to supply chain, manufacturing, retail, healthcare, and document intelligence
- 🔬 Researching multimodal AI, computer vision, bioinformatics, and assistive technologies
- ☁️ Working with AWS, Google Cloud, Docker, and production-oriented ML workflows
- 🌎 Collaborating in Portuguese, English, and Spanish
- 📍 Based in Manaus, Amazonas, Brazil
Supply-chain optimization → Genetic algorithms with a measured 20% improvement
Enterprise document AI → Multi-agent pipelines for extraction, matching, and validation
Retail intelligence → Forecasting and analytics over 1M+ inventory transfers
Energy forecasting → LightGBM pipeline reaching R² 0.9748
Assistive computer vision → Detection, depth estimation, and scene understanding
Bioinformatics → Reproducible 16S rRNA microbiome analysis and ML
| Project | What it does | Core technologies |
|---|---|---|
| FLORA | Local-first pipeline for 16S rRNA microbiome analysis, compositional normalization, and machine learning | Python, DuckDB, CLR/TSS, ML |
| MedVisionFramework | Multimodal medical-image triage and benchmarking using vision-language and foundation models | PyTorch, CLIP, SigLIP, DINOv2, RAG |
| Enterprise Multi-Agent AI | Specialized agents for OCR, cleanup, normalization, matching, enrichment, allocation, and validation | Python, Gemini, agent orchestration, APIs |
| Supply-Chain Optimizer | Allocation engine combining genetic algorithms, optimization baselines, and business constraints | DEAP, Pyomo, Python, analytics |
| Assistive Mobility AI | Computer-vision pipeline for object detection, depth estimation, and scene descriptions | YOLO, MiDaS, BLIP, OpenCV |
| Flashify | AI-powered study platform for generating flashcards, quizzes, and personalized learning paths | Python, TypeScript, PWA, generative AI |
- Production-grade AI agents and multi-agent orchestration
- Multimodal models and vision-language systems
- Scalable ML architectures and distributed data processing
- Operations research and AI-driven decision systems
- Bioinformatics, digital health, and assistive technology
I'm open to collaborating on Applied AI, Machine Learning, Data Science, optimization, and research projects.
[](https://linkedin.com/in/joão-pedro-pereira-santiago-340048269)
[](mailto:jpedropsss@gmail.com)
Building intelligent systems from the Amazon to the world 🌎