I’m Mark Bartolome. I build developer content, and the systems that ship it.

I started out as a software engineer. Spent sixteen years at Goldman Sachs — in as a summer intern, out as a vice president, the last eight of them in senior developer roles — writing Java, C++, and .NET for trading platforms and compliance systems, eventually leading a team of engineers building a Federal Reserve-mandated AML application. Somewhere along the way I figured out the part of the job that resonated with me the most — as much as shipping the code — was making the code legible to the next person who had to use it.

I jumped to AWS as a programmer writer in 2018, where I owned the public technical content for Amazon QLDB across private beta, public preview, and GA: Developer Guide, API references across every supported SDK, the Console UI text, and the help panel copy. That was my crash course in what serious developer documentation looks like at scale: ~21K monthly visits, every example tested, every release note coordinated across product, engineering, and marketing. I created code lab curriculum for the QLDB workshops at AWS re:Invent 2019. Eventually I moved into management, owning Amazon Managed Blockchain and Amazon FinSpace docs.

At Snowflake I led documentation for the Developer and AI/ML surface: Snowpark, the Python/REST APIs, CLI, Cortex AI, Snowflake Intelligence, and Snowflake ML. I split my time between the team — hiring, calibration, editorial review, roadmap — and the systems that made the team faster: an automated test framework that ran every in-doc code sample against the live API, a Python + Sphinx pipeline that generated reference and release notes from source on every API release, and team-wide adoption of AI-native authoring workflows (Claude Code, Cursor, MCP).

I’m at Braintrust now, as a documentation engineer on a three-person team. Braintrust is an AI/agent observability and evaluation platform. Teams at Notion, Stripe, Zapier, Vercel, and Ramp use it to compare models, test prompts, and catch regressions. I cover the Brainstore query engine, enterprise and billing, and the core platform surfaces, and the work runs in two directions. One is rebuilding and optimizing sections that had outgrown their shape: access control, human review, experiments, the SQL reference, and most recently prompts, where two long pages that had drifted apart became one section that’s easier to navigate. The other is authoring content for new features or initiatives, from the Query API reference to the launch docs for SQL charts’ public preview to a cost-optimization guide behind a customer email campaign. Alongside both, I help build the infrastructure the team writes on: Mintlify, agent skills, GitHub Actions workflows, automated gap analysis, the style conventions we hold each other to, and docs evals that score how well agents answer customer questions from our docs.

The thread across all of it: I write better docs because I still read the code, and I lead better teams because I’ve been the IC writer on the team. The next phase of developer documentation is about structured content that serves both humans and agents, and at a company whose product is evaluating AI systems, that stopped being a forecast. It’s the daily work.