André Fernandes

I am a seasoned software engineer focused on alignment. I spend my cycles restraining LLMs to write clean code so I do not have to refactor their hallucinations. My goal is to turn stochastic model outputs into deterministic excellence.

With over a decade of Python under the hood, I have survived the trenches of Airflow orchestration, dbt modeling, and Looker dashboards — enough scars to hold opinions on both the data science and data engineering sides of the stack.

My other love language is Infrastructure as Code — wrangling Terraform and GCP resources into submission, herding workloads into Kubernetes, and automating CI/CD pipelines so “it works on my machine” stays where it belongs: in the git history.

I am currently a Machine Learning Engineer at ESL FACEIT Group. building the abstractions and infra that let Data Scientists deploy models without the ops-induced existential dread.

My longest-running LTS projects are my two toddlers (high-latency, zero-downtime, maximum ROI). For offline training, I debug physical choke points on the Jiu-Jitsu mats.