L'Optimist
Fractional Chief of AI for SMB.
Turning back-office busywork into systems that quietly run themselves.
Fractional Chief of AI for SMB.
Turning back-office busywork into systems that quietly run themselves.
I'm a Fractional Chief of AI for small and medium businesses. I work with managers and CEOs to figure out what's actually worth building with AI, then I build it.
The advising and the building go together. I'm in the room with the executive team for the strategy, with the IT lead for the constraints, and with the people who'll use the tool to make sure they actually use it.
Recent work spans medical device manufacturers and health data companies, where compliance and IP constraints make security a top priority.
Background: data scientist, and engagement manager, with an MS in Applied Mathematics and a BA from the University of Chicago.
Discovery-first assessment of where AI and automation actually help, and where they don't. Build-versus-buy recommendations, vendor evaluation, and a roadmap matched to what the team will actually use.
The repetitive back-office work — compliance checks, reporting, data QA, project-management busywork — rebuilt as systems that run themselves. I build the tools the strategy calls for and stay close to the people who'll use them, so the work gets adopted and the busywork quietly disappears.
Built an FDA recall-monitoring tool that scrapes government databases, matches recalls against the client's product lines, and feeds results into their audit submissions. Work that used to take one intern three months now happens in the space of a click. Delivered in a month, it's been in daily use by the compliance group ever since. Built with Claude Code, with no AI model in the production loop, which mattered for the client's IP and regulatory constraints.
Claude Code, AzureBuilt a set of shareable Claude Skills that standardized project-management work across the entire organization. One runs a QA check on every dataset before delivery, catching DUA and HIPAA violations before they ever reach a client. For this company that's existential: a single data leak could cost them their Data Use Agreement with CMS — the basis of their data access and their whole business model. Another skill takes over repetitive PM tasks on Monday.com. Built to be shared and reused company-wide for enterprise-level standardization.
Claude Cowork, MCPBuilt two tools for a technology consulting firm, both focused on the quality of client deliverables. One automated routine report generation, making deliverables faster to produce and more consistent. The other is a web application that lets the firm's engineers check their machine learning models for bias in the outputs.
PythonIf you're weighing an AI project and want an honest read on whether it's worth doing, get in touch.
contact@l-optimist.com