2
2
Tunisia
10+ years of experience
I am Oussema Toumi, an AI & Automation Strategist, Cybersecurity Specialist, Business Developer, and technology entrepreneur based in Tunisia. I work at the intersection of artificial intelligence, automation, cybersecurity, software architecture, and business development, building intelligent, secure, and scalable digital solutions. I have been working in IT and technology since 2008, with expertise spanning cybersecurity, software development, cloud computing, AI, automation, Linux, networking, and digital transformation. I work with technologies including Google Cloud, Vertex AI, open-source LLMs, Vercel, n8n, Make, and Hugging Face, supported by professional certifications in cybersecurity, networking, Linux, Salesforce, analytics, and digital marketing. I am the creator and architect of projects including Spectral, an AI-powered precision-agriculture ecosystem combining satellite imagery, multispectral drones, IoT, AI, cloud computing, and blockchain; Immo 3.0, a technology-driven real-estate ecosystem; and Medina Experience Live, a digital tourism and cultural experience marketplace. I also build AI agents and autonomous automation systems for research, software development, cybersecurity, business development, data analysis, and operational workflows. My approach is to connect AI with real business processes rather than treating it simply as a conversational technology. I work extensively on international business development, strategic partnerships, and technology-driven growth. My combination of technical and commercial expertise allows me to understand both how to build a system and how to create measurable business value from it. My focus is on making technology reduce complexity, improve transparency, automate operations, and create sustainable value. I am building toward intelligent digital ecosystems where AI, automation, cybersecurity, and entrepreneurship converge to solve real-world problems at scale.

The VoxSales AI Voice Agent is a fully integrated, production-grade artificial intelligence voice solution designed to connect real-time speech-to-speech AI with a live CRM automation pipeline. Operating from a single browser-accessible cockpit, the system allows teams to conduct instant, bilingual voice conversations with prospective clients. The agent is powered by AssemblyAI and utilizes their highest-fidelity female voice, anna, to ensure clear audio with an optimal dynamic range and zero DC offset. The platform is built for speed, featuring a conversational latency of approximately 300 milliseconds and an instant semantic barge-in latency of under 200 milliseconds, which allows users to naturally interrupt the AI just like in a normal human conversation. During active calls, the agent can execute real-time function tools synchronously without blocking the audio flow. It can smoothly schedule product demonstrations, instantly dispatch Telegram alerts to the SecOps Lead, and capture lead data to update the live intelligence dashboard on the browser cockpit. Once a call concludes, the backend automation seamlessly takes over. Using an n8n workflow, the system feeds the call transcript to the Mistral-large-latest model to perform a complete BANT analysis, evaluating the prospect's Budget, Authority, Need, and Timeline. This intelligence is automatically logged into an Airtable CRM, and a final scorecard is sent directly to the team via Telegram, ensuring all data is perfectly synchronized without manual data entry.
30 Sep 2026

VanguardOS is an autonomous AI trading agent built around Alpaca, designed to combine intelligent market analysis, automated decision-making, risk management, and continuous learning into a single auditable system. It uses a multi-brain architecture that combines specialized trading intelligence with ensemble decision-making and a Brain Teacher for shadow trading and model improvement. VanguardOS analyzes real market data across multiple timeframes, evaluates market conditions and sentiment, applies deterministic risk controls and circuit breakers, and generates trade proposals before execution. Its execution layer operates in Alpaca PAPER mode, allowing the system to demonstrate autonomous trading behavior without exposing real capital. The platform also integrates real-time market streaming, portfolio reconciliation, automated scheduling, memory, options capabilities, MCP tools, and a web dashboard for monitoring decisions, positions, trades, and system performance. A key focus of VanguardOS is transparency: paper execution, shadow trading, backtesting, and synthetic testing are kept conceptually separate so performance can be evaluated against real broker data rather than simulated claims. The goal is to demonstrate how autonomous AI agents can operate safely, continuously, and measurably in a real-world financial environment.
4 Sep 2026