AI Research Engineer · AI Systems · Scientific ML
São Paulo, Brazil · EU Citizen (Italy)
I build intelligent systems under uncertainty.
My work sits at the intersection of AI systems engineering, Scientific Machine Learning, and dynamical inference: multi-agent platforms, retrieval infrastructure, evaluation pipelines, probabilistic modeling, and interpretable learning systems for environments we cannot fully observe.
During the day I build production AI systems.
At night I ask whether their dynamics are even identifiable.
- AI systems engineering and agentic infrastructure
- Multi-agent orchestration, MCP, retrieval, and tool execution
- Evaluation, observability, and reliability for LLM-based workflows
- Scientific ML for dynamical systems under uncertainty
- Probabilistic inference, regime transitions, and survivability under perturbation
- Research software engineering for uncertainty-aware intelligent systems
I currently work as an AI Research Engineer at BTG Pactual, building production AI systems and developer-facing AI tooling.
My engineering work focuses on:
- LLM platforms and internal AI tooling
- multi-agent workflow orchestration
- retrieval infrastructure and context engineering
- evaluation and observability for agentic systems
- human-in-the-loop validation and controlled generation
- reusable AI artifacts, tools, and workflow systems
Some internal details are intentionally omitted. Enterprise systems are, tragically, not open-source playgrounds for résumé dopamine.
Founder of NOVA Research Systems — independent research on AI systems under uncertainty.
Local survivability estimator for dynamical systems under perturbation.
HSP measures how much of the local basin geometry remains accessible under uncertainty:
Current work explores neural approximations via Continuous Flow Matching, probabilistic latent dynamics, and uncertainty-aware dynamical inference.
Interpretable regime discovery in dynamical systems.
Combines Bayesian online change point detection, sparse regression, uncertainty estimation, and Pareto-based model selection to identify governing equations across regimes.
Applied to infrastructure monitoring and early degradation detection.
Rare-event ecological modeling under uncertainty.
Neural State Space Models and Deep Markov Models for survival analysis, distribution shift, and counterfactual ecological dynamics.
Engineering Projects
Humainze
Integrated intelligent IoT and AI platform.
Built across embedded systems, backend services, ML pipelines, and cloud deployment:
C++ firmware for ESP32/Arduino sensor ingestion
Java/Spring Boot backend services
Python ML module for forecasting and drift detection
Streamlit dashboard for monitoring and interaction
Docker-based deployment on Azure infrastructure
Related repositories:
humainze-iot
humainze-ai
MIDAS Java
Enterprise backend engineering project built with Java and the Spring ecosystem.
Focused on production-oriented backend patterns:
Spring Boot and Spring Security
JWT and OAuth2 authentication flows
JMS-based asynchronous messaging
Flyway migrations
Oracle DB integration
modular REST API design
Ceci
LLM-based system with RAG, MCP integration, and structured guardrails.
Awarded 3rd Place at Next 2025 for system design and robustness after evaluation by internal and external technical panels.
Recognition
1st Place — Global Solution 2025 · Extreme event forecasting using LSTM + Attention
3rd Place — Next 2025 · Ceci · LLM-first system with structured guardrails
Stack
AI Systems
LangChain · LangGraph · Agno · LiteLLM · MCP · RAG · Langfuse · Multi-agent Systems · Evaluation Pipelines · Observability · DAG Orchestration · Context Engineering
Scientific ML
JAX · Equinox · Diffrax · PyTorch · PyTorch Geometric · Neural ODEs · Neural SDEs · Neural SSMs · PINNs · SINDy · Flow Matching · Bayesian Inference · Monte Carlo Methods
Engineering & Infrastructure
Python · Java · C++ · TypeScript · R · Julia · FastAPI · Spring Boot · Docker · Kubernetes · MLflow · ONNX · Neo4j · FAISS · Linux
Cloud & Compute
AWS · Azure · GCP / Vertex AI · CUDA · Triton
Interested in research and engineering environments working at the intersection of:
AI Systems, Scientific ML, dynamical systems, probabilistic inference, and AI for science.
Currently focused on building interpretable, reliable, and uncertainty-aware intelligent systems.
EU Citizen · Open to relocation