Contact person: Business Client

Phone:Show

Email:Show

Location: Winslow, United States

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"Project Overview

We are building a production-grade Solana trading system that combines:

A deterministic Rust execution core

A strict, finite Action API

A semi-autonomous AI decision layer that selects from allowed actions only

This is not an experimental bot, retail toy, or prompt-driven system.

The AI never:

holds private keys

signs transactions

sets raw trade sizes

bypasses limits

All risk controls are enforced below the AI in Rust.

This project is already architected. You will be implementing defined components, not inventing the system.

---

What You Will Be Working On

You will help implement and wire together:

Action schema validation (strict JSON, fail-closed)

Strategy modules (bounded, deterministic, profile-based)

Telemetry aggregation (sanitized summaries only)

AI decision loop (single action per cycle, no chaining)

Simulation / backtesting mode (no RPC, no signing)

Logging and audit trails (decision → action → outcome)

You will not be responsible for key management, signing logic, or live fund custody.

---

What You Will NOT Touch (Non-Negotiable)

You will not work on:

Private key handling

Transaction signing

Withdrawals or fund movement logic

RPC credentials tied to funds

Risk limit enforcement logic

Any system that allows AI to invent actions or parameters

Any attempt to expand scope into these areas is a hard stop.

---

Required Skills & Experience

Must Have

Strong Rust experience (systems-level, async, state management)

Familiarity with Solana concepts (transactions, RPC, execution flow)

Experience designing deterministic, fault-tolerant systems

Comfort with strict schemas, validation, and defensive coding

Strong Plus

Experience with trading systems or automation

Understanding of risk isolation and failure modes

Experience integrating AI systems without trusting them

---

What We Care About

We care about:

determinism over cleverness

bounded risk over flexibility

explainability over opacity

systems that fail safely


We do not care about:

hype

“AI magic”

over-engineering

unnecessary complexity

---

How Success Is Measured

You are successful if:

The AI can go offline and funds remain safe

Invalid AI outputs are rejected cleanly

Strategies cannot exceed defined limits

Every action is traceable and explainable

Simulation behaves identically to production logic

---

Engagement Details

Scope: Medium-term (milestone-based)

Start: Immediate

Collaboration: Async, documentation-first

Access: Code only (no keys, no wallets)

---

How to Apply

Please include:

1. A brief description of a safety-critical system you’ve worked on

2. Your experience with Rust and Solana

3. How you think about risk boundaries in automated systems

If your answer involves “just trusting the AI” or “prompting it better,” this is not a fit." (client-provided description)