Zakaria ABIDI@zakariaabidi47525
2
Events attended
1
Submissions made
France
10+ years of experience
About me
I'm a CTO, software engineer and cloud architect with nearly 20 years of experience designing, building and scaling software platforms across multiple industries. My background spans software engineering, cloud infrastructure, DevOps, AI-powered products, geospatial systems, data platforms, automation and SaaS architecture. Throughout my career, I've worked across the full technology stack, from backend development and infrastructure design to product strategy and technical leadership. I enjoy solving real-world problems through technology and turning ideas into scalable products with tangible business value. For this hackathon, I'm focused on building practical AI solutions that go beyond demos and showcase how AI can deliver meaningful impact in everyday life.
🤓 Latest Submissions
ZikosoftTrader AI
ZikosoftTrader AI is a risk-governed multi-agent trading platform built for Alpaca Paper Options. Its core principle is simple: AI can analyze, propose, criticize, and explain, but it cannot bypass deterministic risk controls. The system uses four specialized agents. The Market Agent gathers market data, news, OHLCV, snapshots, and options evidence through Alpaca and the official MCP server. The Strategy Agent evaluates three strategies—AI Market Agent, Moving Average Crossover, and RSI Reversal—and converts qualified directional signals into conservative long call or long put candidates. The Risk Critic challenges each proposal, while a separate deterministic Risk Engine applies hard controls such as data freshness, position limits, option DTE, spread, premium, buying power, duplicate protection, human approval policy, and a global kill switch. Only approved decisions reach the dedicated Order Worker, which submits paper orders to Alpaca and tracks trade updates. The platform also includes an Agent Room for full decision transparency, Ask Ziko for natural-language explanations, Historical Replay for deterministic demos and backtesting, portfolio monitoring, AI budget controls, system health monitoring, and a multilingual React dashboard. The result is an explainable, auditable, paper-only architecture designed to demonstrate safer AI-assisted trading without giving an LLM unrestricted execution authority.
4 Sep 2026
