Inspiration

What it does

How we built it

Inspiration

Field operations generate a surprising amount of evidence: timestamps, GPS positions, tracking trails, photos, project context, and activity history.

LogDelDia already captured this evidence as part of real workforce operations. The problem was what happened next.

As the number of activities grows, a supervisor cannot realistically inspect every GPS trail, tracking gap, timestamp, and historical pattern. We saw an opportunity for AI—not to replace the supervisor, but to investigate the evidence first and help humans focus their attention where it matters.

That became LogDelDia AI.

What it does

LogDelDia AI provides evidence-grounded AI supervision for field operations.

A supervisor can select an existing field activity and request an AI review. LogDelDia retrieves and computes operational evidence including activity duration, GPS tracking quality, temporal coverage, tracking gaps, worker history, and project context.

This structured evidence is analyzed using NVIDIA Nemotron through Nebius Token Factory.

The result is not simply a generic LLM response. LogDelDia returns structured findings, risk, missing evidence, and a recommendation that the supervisor can evaluate.

Importantly, the system distinguishes between an anomaly and proof of wrongdoing.

For example, during our real-world testing, the system detected a tracking gap that exceeded the configured threshold. However, it also recognized that there was insufficient worker history to establish a reliable baseline. Rather than inventing a cause or making an accusation, it explicitly reported

DEMO INSTRUCTIONS: https://logdeldia.com Judge account username: judge_supervisor Password: "Judge123" The account is configured as a supervisor and has access only to a dedicated demo worker and demo activities.
To test LogDelDia AI:

  1. Sign in with the credentials above.
  2. Open the demo worker's activity history.
  3. Select a completed activity.
  4. Click “Revisión IA”.
  5. LogDelDia will assemble verified operational evidence and send it to NVIDIA Nemotron through Nebius Token Factory.
  6. Review the resulting risk assessment, findings, missing evidence, and recommended follow-up.

The AI analysis is advisory. The final decision always remains with the human supervisor.

Built With

Share this project:

Updates

Submission history