Project Seal India Task Specialist

I’m running a series of tightly-scoped tasks under the Seal India initiative and need a partner who can take each assignment from prompt design through final submission without hand-holding. Every task pays $30 on successful delivery, and the quicker you can turn them around, the more I can queue up for you.

You’ll be working with large-language models—ChatGPT-5.5, Claude, or any equivalent you already subscribe to—probing them for search and reasoning failures. I supply the overall objective; you create the cleverly-crafted prompts, document outcomes, and package the results exactly as the task spec requires.

My ideal collaborator already:

• Uses advanced AI tools daily and is comfortable chaining prompts, applying system messages, and analysing token-level behaviour.
• Runs a reliable residential proxy so IP rotation or geo-targeting never slows you down.
• Has previous hands-on experience with model debugging, AI model training, and AI-powered web scraping, so edge-case thinking comes naturally.

Deliverables for each task
1. A brief rationale for your prompt strategy.
2. The full prompt set (system, user, follow-ups) with any temperature / max-token tweaks noted.
3. Raw model output plus a concise analysis highlighting failure modes.
4. A final, well-formatted submission ready for my portal.

Acceptance criteria
• The model failure exposed must match the objective I provide.
• All prompts and outputs must reproduce on my end when I validate them behind a different proxy.
• Turnaround per task: 24 hours or less.

If this workflow sounds like second nature to you, let’s start with the first assignment immediately and keep the pipeline moving.