GS / ML SYSTEMS
PROFILE: ACTIVE FOCUS: ML SYSTEMS BUILD: ITERATING

About

Gilberto Silva — Data Systems & Applied ML

Back to Portfolio
Profile

I build end-to-end data systems that actually run in production.

My work sits at the intersection of data engineering, machine learning, and infrastructure: designing pipelines that move from raw ingestion to reliable, monitored inference. I care about systems that do not just work once, but keep working: versioned data, reproducible training, observable services, and models that survive real-world usage.

Background in biomedical engineering and research, now applied in industry solving high-volume, real-time data problems. I have worked across the full lifecycle: data collection, modeling, deployment, and feedback loops, with a strong bias toward practical impact over academic elegance.

This space is a hub for my work: experiments, systems, and projects bridging research and production.

Core Areas
Real-time & streaming data systems ML lifecycle: training to monitoring Experimentation, tracking & reproducibility Data modeling, contracts & governance Observability and production reliability Edge data collection: IoT / LoRa / sensors
Beyond Core ML

IoT and edge systems: LoRa, sensors, distributed data collection

Data visualization: turning raw signals into decisions

Independent products: building data-driven tools and platforms

Photography: a slower way of observing systems in the real world

Research & Publications

Research remains part of the foundation: evidence, method, and disciplined iteration. My publication history sits mostly in the biomedical engineering and applied research space, and it continues to shape how I build production systems today.

Biomedical engineering foundation

Grounded early work in experimental design, measurement discipline, and translating data into usable conclusions.

Applied research mindset

Reinforced the habit of reproducibility, traceability, and validating results instead of trusting intuition.

Production impact today

That same rigor now shows up in lifecycle-aware ML systems, governed data flows, and deployable inference services.

Tooling Philosophy
I favor simple, composable systems over complex abstractions.
Postgres over distributed systems until necessary.
Observability before optimization.
Reproducibility over speed in early stages.