Mind the Gap, Then Measure It

Description and use

A Tinker LoRA fine‑tuned model turns vague accessibility complaints into actionable work orders by providing the exact measurement needed, closing the gap between reporting and fixing barriers.

Problem

City residents report accessibility barriers but lack the specific evidence needed for public‑works crews to act, causing delays and unaddressed hazards.

Outcome

The system turns vague complaints into actionable work orders by providing the exact measurement needed, reducing the gap between reporting and fixing accessibility barriers.

Workflow

• 1. Resident copies a messy field note into the Flask app.
• 2. The app sends the note to the Tinker fine‑tuned model.
• 3. Model outputs structured JSON with missing evidence and a concrete measurement instruction.
• 4. Resident follows the instruction, measures the barrier on site.
• 5. Resident submits the measurement back to the city, completing the work order.

Evidence

The model returns a structured civic report and the single most useful measurement you still have to go take: "missing: the vertical lip height at the bottom of the ramp. Go measure it in inches; anything over 1/4 in is a barrier."

Source

In all three, the model correctly identifies that the resident described the problem but not the dimension a crew needs, and it hands back a single, concrete thing to go measure.

Source

The model does not echo the real address. It hallucinates a plausible‑looking Cambridge address… The product’s answer to "the model guessed the address" is the same as its answer to any missing field: a human verifies it on the ground.

Source