Samuele Giampieri

I'm the creator of RedAmon, an open-source platform that automates the work of an entire offensive security team. An autonomous AI agent moves through the same stages a human red team would, gathering intelligence, attempting to break in, then probing deeper once inside, while a person stays in control and approves every sensitive action. Its standout feature is that it doesn't just find vulnerabilities; it fixes them. Confirmed issues are ranked by real-world risk, and the system opens a ready-to-review fix as a pull request directly on the client's code. It can run several specialist agents in parallel, remembers what it learned from earlier engagements, and draws on more than 400 AI models from a dozen providers to always use the best one for the task.

Security is built into the foundation, with strict isolation and hard limits that make it impossible to aim at protected targets such as government systems, and the whole platform installs with a single command. RedAmon has grown into a thriving open-source project with over 2,200 GitHub stars, more than 200 daily downloads, and an active community of contributors, and is used by enterprises and security professionals worldwide.

pMag is a white-label platform that lets a company turn any internal capability into an AI-powered tool in minutes, not months. Once a business function is connected, it instantly becomes something a conversational AI assistant can run on request, schedule to run automatically, and monitor, with no rebuilding required. Because it is built to be multi-tenant, one shared core quietly powers nine tailored products across six different industries, so each client gets a bespoke experience while only a single codebase is maintained. Real deployments show its range: a banking system that reviews financial guarantees, checks issuers against official registers, spots forged logos and tampered documents, and returns a simple red, yellow, or green fraud verdict; and a multilingual order-entry system that reads incoming orders, validates them against business rules, and files them straight into the company's SAP system. It can tap several leading AI providers and automatically choose the most cost-effective one, and it runs equally well on a single server or scaled across the cloud as demand grows.

SMA is an enterprise assistant that lets employees get trusted answers from their company's document library by simply asking. Instead of hunting through SharePoint for a policy, a past project, or a procedure, staff ask a question in plain language and get a grounded answer with links to the exact source, right down to the precise moment inside a recorded meeting when it is relevant. Behind that simplicity is a team of specialized AI agents, each expert in a different area such as HR policy, project history, or navigating shared folders, coordinated by a controller that always routes a question to the right one.

The system automatically keeps itself up to date, picking up new and edited documents as they are added, so answers never go stale. It was built for real corporate use: single sign-on, strict access controls so people see only what they are allowed to, and enough efficiency to serve hundreds of employees at once on modest infrastructure. The result replaces slow, frustrating internal searches with fast, reliable, source-backed answers.

Pathbreak is a SaaS Cloud security platform that answers the question most tools can't: not which weaknesses exist in theory, but which ones an attacker could actually chain together to reach a company's most valuable data. It builds a live map of a company's cloud and then proves, step by step, the real routes an attacker could take, from an exposed server, through borrowed permissions, all the way to a sensitive database, so teams stop drowning in thousands of hypothetical alerts and focus on the few paths that are genuinely walkable.

Every step comes with clear evidence, so findings can be trusted instead of second-guessed. Once the real attack paths are known, Pathbreak ranks the fixes that shut down the most routes at once, and a built-in AI assistant explains each path and its remediation in plain language. New attack techniques can be added as a simple configuration rather than new code, keeping it current as threats evolve. The result turns cloud security from a noisy checklist into a focused, prioritized action plan.

Reranka is a custom business intelligence platform that provides an online retailer with a single, real-time view across more than a dozen marketplaces, including Amazon, eBay, and Zalando. It continuously tracks tens of thousands of products every hour and turns a flood of raw sales data into clear dashboards for revenue, stock health, and profitability, all the way down to the level of individual products. Its most valuable feature is automated price optimization: the platform tests prices on live listings, monitors how sales respond, and learns the exact price that maximizes profit for each product type, taking the guesswork out of pricing. It also spots which products customers tend to buy together so they can be bundled, forecasts demand to avoid costly stockouts, and helps win the all-important Buy Box.

Everything is presented through visual, easy-to-read reports with one-click export, so business teams, not just analysts, can act on it. The platform handles enormous volumes of data reliably and refreshes throughout the day, giving decision-makers an always-current picture of their market.