Hi, I’m Sreenath R.

I build local-first AI tools and transparent motorsport strategy software. I’m currently an ESSEC MiM 2026 student, developing practical Python systems around endurance racing, telemetry, and decision support.

Pitwall Agent is installable endurance-strategy software with:

• deterministic fuel, stint, pit-service, tyre, and driver-rule tools;
• editable race configuration and telemetry-quality auditing;
• seeded strategy comparison and explicit evidence levels;
• optional local Ollama tool calling;
• a full no-model mode with JSON automation and pit-sheet export.

git clone https://github.com/Dabi-init/endurance-stint-planner.git
Set-Location .\endurance-stint-planner
powershell -ExecutionPolicy Bypass -File .\run.ps1

Read the project · Try the alpha release · Join the discussion

• publishing the first anonymised predicted-versus-actual validation case;
• adding legal-to-share simulator telemetry adapters;
• making strategy change triggers explicit and testable;
• exploring read-only MCP interoperability for the audited race tools.

• deterministic calculations underneath language models;
• evidence and uncertainty before confident claims;
• local data boundaries by default;
• reproducible examples, adversarial tests, and honest limitations.

Feedback, non-confidential validation data, and focused contributions are welcome through the Pitwall Agent issues and Discussions.