Felipe Reis — Senior ML & AI Engineer

Hi, I'm Felipe. From experiment to production — at any scale.

6+ years building production ML and AI systems across startups and enterprise. AWS-certified ML Engineer with deep expertise in Kubernetes, LLMs, and MLOps. Driven by clean code, solid infrastructure, and measurable impact.

Online

Stack

Python Go AWS Kubernetes LLMs Terraform

Technical Focus

Machine Learning Systems

Designing, training, and productionizing models across the full lifecycle — from experimentation to reliable, low-latency inference at scale. Built platforms serving tens of millions of predictions per week.

AI Engineering

Building LLM-powered applications and chatbots, RAG pipelines, and AI agents. Instrumenting and monitoring AI systems for quality, drift, and cost in production environments.

MLOps & Deployment

Containerized training workflows, experiment tracking with MLflow, automated CI/CD pipelines, and model versioning for reproducible, auditable model delivery on SageMaker and Kubernetes.

Data & Feature Pipelines

Building robust ingestion and transformation pipelines with Spark and Airflow. Designing feature stores — both online and offline — that keep training and serving distributions consistent.

Observability & Monitoring

Instrumenting models and infra with Prometheus and Grafana. Detecting data drift, scoring degradation, and chatbot quality regressions through automated alerting and structured dashboards.

Platform Engineering

Architecting cloud-native ML platforms on AWS — compute clusters, model registries, Terraform-managed infrastructure, and the Kubernetes backbone that ties experimentation, training, and serving together.

Stack

Python

Scala/Spark

AWS

Kubernetes

Terraform

MLflow

SageMaker

Kubeflow

Airflow

Prometheus

Grafana

PyTorch

LightGBM

LLMs / RAG

FastAPI

Docker

Kafka

Argo

CI/CD

Building Production ML Systems: Lessons from the Field

Key insights from building and deploying machine learning systems at scale.

Read

How I Work

Production first

I think about serving, monitoring, and rollback from day one — not as afterthoughts.

Less complexity

A simpler model that ships beats a sophisticated one that doesn't. I prefer the minimal solution that achieves the goal.

Ownership

I stay involved through deployment and post-launch. Experiments need to land, not just run.

Hi, I'm Felipe. From experiment to production — at any scale.

Technical Focus

Machine Learning Systems

AI Engineering

MLOps & Deployment

Data & Feature Pipelines

Observability & Monitoring

Platform Engineering

Stack

Courses

Recent posts

Building Production ML Systems: Lessons from the Field

How I Work

Production first

Less complexity

Ownership

Open to new projects