Stenio Ferreira

The Reliability Guy

Stenio Ferreira's job title has changed six times since 2016, but the job itself hasn't moved much. At HashiCorp it was Vault - convincing companies to trust a piece of software with their passwords, API keys, and certificates instead of a spreadsheet. At Google it was risk and compliance tooling. At Grafana Labs it was observability, the dashboards that tell you something is on fire before your customers do. Now, at Temporal Technologies, it's durable execution - workflows and AI agents built so that a crash, a timeout, or a flaky API call doesn't erase their progress and force them to start over.

He works out of New York as a Staff Solutions Architect, which in practice means he sits across the table from engineering teams and explains, in whatever language gets through, why a system that remembers its own state is worth the switching cost. His GitHub account is a quieter version of the same résumé: a durable-agent-loop reference implementation, a Temporal-and-Nexus cross-namespace demo, a project comparing Google's Agent Development Kit against Temporal head to head. Not marketing copy. Working code, published, with commit histories.

The path here started further back than Temporal or Grafana. Ferreira studied Computer Science at UEL, a public university in Londrina, Brazil, then crossed into business with a part-time MBA at the University of Chicago's Booth School, finishing in 2010. At Booth he sat on the MBA Admissions Committee and co-chaired the school's Innovation Group - early evidence of the same instinct that shows up later in his career: technical competence paired with an appetite for explaining it to people who don't share the vocabulary.

That combination took him to Slalom first, where he worked as a Solution Architect and published some of his earliest technical writing - a 2016 piece on secret management, a 2019 walkthrough of Terraform Enterprise integrations with ServiceNow. Then HashiCorp, where he stayed more than five years, eventually becoming LATAM Senior Solutions Engineer. He is credited with building the first phase of HashiCorp's Latin American market - the unglamorous work of demos, procurement conversations, and technical trust-building across a region and several languages. In 2020 he published a formal HashiCorp Vault performance benchmark, the kind of document sales engineers write when they want their claims to survive scrutiny.

From there: Google, as Global Solutions As Code Manager for Risk & Compliance, then Grafana Labs as a Staff Solutions Engineer covering the New York metro area. Somewhere in the middle of all this he accumulated a stack of certifications that would take most people a decade to earn on top of a full-time job - AWS Certified Solutions Architect Professional, AWS Security Specialty, AWS Machine Learning Specialty, AWS Data Engineer Associate, and four separate Google Cloud Professional certifications covering architecture, data engineering, security, and networking. He also holds a PMP. It reads less like résumé padding and more like someone who treats staying current as a discipline, the way other people treat exercise.

His most recent move is to Temporal Technologies, where the company's core pitch - that software workflows should be durable, able to pick up exactly where they left off no matter what fails underneath them - has found a new urgency now that AI agents are being asked to run multi-step, long-lived tasks instead of answering single questions. Ferreira's public GitHub work leans directly into that problem: a reference implementation for durable agent loops, a demo of Temporal agents talking to each other across namespaces using Nexus, a side-by-side comparison of Google's ADK framework running with and without Temporal underneath it. These aren't polished product demos. They're the kind of exploratory, slightly rough code a solutions architect writes to figure out where a new pattern breaks before a customer does.

Outside the infrastructure work, Ferreira's public footprint stays low-key. His X account, dormant for long stretches, lists interests spanning AI, security, blockchain, and heavy metal, and has existed since June 2008 - a relic of an earlier internet era, carried forward into this one. A 2017 Medium post shows him taking notes in a music theory class at Chicago's Old Town School of Folk Music, evidence of a life that includes an electric guitar somewhere alongside the Terraform state files. He lists Portuguese as a native language, Spanish at full professional fluency, and German at an elementary level - useful, presumably, for a career that has involved building markets and relationships across at least two continents.

What makes Ferreira's career legible isn't any single dramatic move. It's the consistency of the problem he keeps choosing: systems that people depend on without thinking about, until the moment they fail. Vault secrets. Compliance workflows. Observability dashboards. Now, AI agents that need to survive their own mistakes. Each role has been a different vantage point on the same question - what happens when something breaks, and how do you make sure it doesn't cost anyone their data, their money, or their time. It's not a founder's story. It's a practitioner's, told in certifications, commits, and a newsletter nobody asked him to write but that a few thousand executives apparently read anyway.