AI-native quant. Live trading systems + agent ops.
Finance degree, retail trading roots. The gap between a strategy that looks good in Python and a strategy that actually runs in production — broker quirks, OCO failures, terminal crashes, kill governance, multi-agent coordination — is what I kept running into. So I built the whole stack.
Currently building StoneDesk: a prop trading firm with 3 production agents (DeskHead, QD, Researcher) orchestrating live multi-broker execution. Forked paperclip-stonedesk with planexec adapter, phantom-XML strip, and OpenCode local routing. Shipped HaltActor governance, file-bridge EA architecture, GitNexus code-intelligence MCP. All running on real money, on FundedNext / E8 / LiteFinance, since 2026-05.
Available for short-term contracts: agent ops, MT5 / Nautilus integration, prop-firm infra, custom Claude / Ollama tooling. Global engagements welcome — Togolese passport, offshore-friendly.
Quant
- · Nautilus integration
- · MT5 / MQL5 / EAs
- · Prop firm production
- · OCO + hedging brokers
- · Backtesting + risk math
AI
- · Claude / Ollama / MCP
- · Multi-agent orchestration
- · Planner-executor patterns
- · Prompt-injection defense
- · RAG + code intelligence
Infra
- · Wine / MT5 / systemd
- · Supervised service patterns
- · Production debugging
- · File-bridge architectures
Languages
- · Python
- · Node.js / vanilla JS
- · MQL5
- · SQL · Bash
- Live STONEDESK · production prop firm
- Live 7tics.com · this site
Let's talk.
Book a 20-min call to scope a project. Async inquiries by email also welcome.