Content & Communications Lead -- Recidiviz
Recidiviz is the only organization I've found that matches how I actually think about this problem. You build technology that makes the criminal justice system's own data work against mass incarceration-- not by arguing from the outside, but by partnering with the agencies that run the system. That's exactly how I've operated for nine years.
I built Steel Man Resumes inside treatment centers and halfway houses-- as an alum, not a consultant. I built AI governance communications for Justice Beacon Solutions alongside a corrections researcher, not about corrections research. I built coalition proposals for six Milwaukee reentry organizations by learning their constraints first, not pitching my solution first.
My specific skills-- translating behavioral science into human stories, writing for government partners and justice-impacted communities simultaneously, building content infrastructure that scales-- exist because I've spent a decade solving the exact communication challenge Recidiviz faces: making data-driven criminal justice reform feel urgent to the people who can act on it, without ever losing the trust of the people it's supposed to serve. I can't do this work anywhere else because nowhere else is doing this work this way.
Justice Beacon Solutions created a new governance discipline-- Justice Decision Observability-- for AI systems in corrections. The narrative challenge: corrections agencies and defense attorneys are adversaries in court, but JBS needs both of them to trust the same governance documentation.
For corrections leaders, the story had to be protection: "When the next in-custody death happens and AI was involved, can you reconstruct exactly what happened between the alert and the human decision? If not, that's where your liability lives. We document it so you can defend it." For defense attorneys, the story had to be evidence: "The human decision layer in corrections is completely undocumented. Our governance reconstruction creates structured, DOI-registered records under a published standard. Your expert gives an opinion. We give a repeatable methodology."
I navigated this by anchoring the narrative in the one thing both audiences actually want: a documented record of what happened. I didn't soften the message for either side. I made the documentation itself the protagonist-- not the agency, not the attorney. The governance record serves both because it's descriptive, not evaluative. It doesn't say who was right. It says what happened.
I built this narrative across the JBS website (audience-specific /for/ pages), a strategic positioning analysis (five buyer personas stress-tested), a CCJ policy response memo, and a full brand doctrine. The site is live at justicebeaconsolutions.com-- you can see each audience page and judge whether the tension holds.
The uploaded piece is a strategic positioning analysis I wrote for Justice Beacon Solutions titled "JBS Positioning Problem: What's the Deliverable?" JBS had built the technology, published 24 DOI-registered governance standards, and launched a portal-- but every conversation with a potential buyer ended the same way: "OK, but what do I get?"
I diagnosed the root cause: this wasn't a messaging problem. It was a category-creation problem. Fleming invented a field that didn't exist before. There was no existing mental model for a buyer to hang it on. Every buyer mapped JBS onto something they already knew-- audit, consulting, compliance, investigation-- and got confused when it didn't fit.
I concepted the approach: five adversarial buyer personas (The Warden, The Defense Attorney, The County Risk Manager, The AI Vendor, The Internal Skeptic), each stress-testing the positioning from a different angle. I conducted the research, mapping how each persona misunderstands the category. I wrote the full analysis-- root cause diagnosis, six-part solution framework, implementation priority-- and then implemented the recommendations across the live website, including a deliverables-first page structure, one-sentence pitch, and a differentiation table clarifying what JBS is not.
My role was end-to-end: problem identification, diagnostic framework design, strategic writing, and implementation. This is the kind of content work I'd bring to Recidiviz: diagnose why a narrative isn't landing, then fix the infrastructure that produces it.