Ai Jarvis prototype

Project: Voice-Based AI Assistant (Flipping Task Only – $1,000, Milestones)

I am building a voice-based AI assistant prototype inspired by a “British butler” style intelligence.
This phase is strictly focused on ONE task only: flipping (buy → evaluate → list → track).

This is not a vague idea and not a full system.
I am hiring someone who can build clean logic and reliable workflows, not flashy demos.

⸻

Objective (Read Carefully)

Build a working voice-based assistant that can:
• Converse naturally via voice
• Maintain basic memory across conversations
• Own ONE task end-to-end: flipping evaluation and listing support

Anything outside flipping is conversation only, not execution.

⸻

Scope – What You Will Build

Core (Required)
• Voice input → AI reasoning → voice output
• Calm, professional “British butler” tone
• Persistent memory with permission-based saving
(assistant asks before remembering non-essential info)

Flipping Task (PRIMARY)

The assistant must be able to:
• Ask required questions (buy price, condition, platform, demand)
• Calculate:
• expected resale
• profit
• margin
• basic risk
• Give a clear YES / NO recommendation with reasoning
• Generate a ready-to-post product listing
• Track flips with simple status (sourcing / listed / sold)

Example behaviour:

“At this price, sir, the margin is insufficient. I would advise against proceeding.”

⸻

Non-Negotiables
• Milestone-based work only (no full payment upfront)
• Access to the project must be provided after each milestone
• Clear communication (short updates are fine)
• If you disappear, the contract ends immediately
• Overpromising = automatic rejection

⸻

Application Test (IMPORTANT – FILTER)

To apply, answer ONLY these two questions.
Do not copy-paste the job description.
1. How would you structure the flipping decision logic so the assistant can confidently say yes or no?
2. How would you handle memory so the assistant remembers important things but avoids storing everything?

Applications that skip this will be ignored.

⸻
Budget

$1,000 total, split into milestones.
This is a prototype, not a production system.

If Phase 1 is done well, there will be follow-up work.

⸻

Final Note

I value clarity, discipline, and logic over speed or flashy features.
If you enjoy building systems that actually work, this will be a good fit.