Founder · AlfaNest Labs · Saint-Fulgent, France
Security software developer I build secure products that respect your data.
Security software developer and solo founder based in France. Therefore, I focus on encrypted vaults, API security tools, AI agent guardrails, and local-first desktop apps. In addition, everything is designed so the server never sees your plaintext.
Security software developer — portfolio highlights
Below is the AlfaNest Labs product line: seven live systems, each with a documented threat model. Scroll the list — the preview on the right follows each project.
Security software developer — vault and desktop products
PIN + biometric vault for passwords, seed phrases, documents, and files. Because the design is local-first, the core vault never requires a network connection. Moreover, zero-knowledge architecture means the server stores ciphertext only.
Bookkeeping app for French BNC independents. As a result, revenue, expenses, quotes, invoices, URSSAF estimates, and bank reconciliation stay entirely on your machine. In conclusion: one-time payment, no subscription, no cloud ledger.
Send invoices, track payments, and stay compliant. Facta-Z is the unified workspace for freelancers and small companies to manage customer records, draft quotes/invoices with live PDF previews, collect online payments via Stripe Connect, and handle electronic invoicing. Designed with compliance in mind, featuring signed PDFs and 10-year encrypted archiving.
Security software developer — API scanning, agents, and machine identity
Paste an OpenAPI 3.x spec or fetch a public HTTPS URL. Subsequently, you get a risk score and findings focused on auth posture and sensitive route exposure — not a full pentest, but actionable signal in seconds.
Decision API that intercepts tool calls before side effects run. For instance, you can enforce allow, deny, or human-approval policies per action type. Furthermore, an append-only audit trail supports governance for teams running LLM agents in production.
Inventory non-human identities: service accounts, API keys, tokens, and certificates. To clarify, client-side AES-256-GCM means the server stores ciphertext. Additionally, HMAC chain integrity and Shamir 3-of-5 recovery ship on Pro.
Security software developer — experimental marketplace
Marketplace for abandoned repos and dead side projects. First, sellers list inventory; then buyers inspect a preview and commit Scraps through Vault escrow. Finally, automated S.P.E.C.I.A.L. scoring by The Overseer keeps listings comparable.
Security software developer — privacy-first work at AlfaNest Labs
I'm Marian, a security software developer and founder based in Vendée, France. Firstly, I started AlfaNest Labs as a micro-entreprise to build tools where encryption happens on your device, not on a server you cannot inspect.
How I approach every product
My approach is straightforward: understand the threat model first, then design the architecture, then build. However, if the answer to “what happens when the server is compromised?” is that user data becomes exposed, I restart the design. Consequently, every release pushes risk left.
Principles I don't compromise on
Beyond security, I care about practical UX and honest pricing. Similarly, I avoid dark patterns and subscription traps on tools that do not need them. Above all, I do not claim certifications I have not earned.
- nameMarian George Grosu
- companyAlfaNest Labs (EI, France)
- locationSaint-Fulgent, 85250 Vendée
- SIREN103 036 695
- SIRET103 036 695 000 15
- NAF62.01Z — computer programming
- emailcontact@alfanestlabs.com
- securitysecurity@alfanestlabs.com
- X@alfanestlabs
- LinkedInalfanest-labs
- Telegram@alfanestlabs
Security software developer — lab notes and experiments
Smart systems require strict, verifiable boundaries. At AlfaNest Labs, we build AI, security, and decision infrastructure under a zero-trust model: every input is verified, every agent action is governed, and every execution boundary is locked by design.
Threat modeling as foundation
We prioritize rigorous threat modeling over rapid, unverified code. Before any system is built, we map failure modes, data leakage vectors, and authorization coverage to ensure the system remains resilient even if individual components are compromised.
User sovereignty by design
Our client-side encryption architectures mean we explicitly design systems so the server cannot access your plaintext secrets. By eliminating central points of failure, we protect your operations from both external threats and vendor compromises.
The EU CRA (effective September 2026) is not a certification wall for micro-operators. Instead, it asks for concrete, public artefacts.
What you actually ship
- A public security policy (
security.txtper RFC 9116) - A coordinated vulnerability disclosure process
- A software bill of materials (SBOM) in CycloneDX or SPDX format
- Security headers and documented encryption choices
- Incident reporting to the relevant national CSIRT within 24/72h if a vulnerability is actively exploited
At AlfaNest Labs, we build products with this compliance built-in. SafeKey and EI-BNC Compta already publish these public-facing artifacts, while our security tools—Clearance, API Risk Monitor, and Agent Security—help teams scan OpenAPI specifications, intercept agent transactions, and map compliance readiness across MiCA, ISO 27001, and SOC 2.
To be announced...
Security software developer — skills and delivery stack
How this security software developer ships
Below is the practical stack behind AlfaNest Labs. In short, cryptography, TypeScript, and operator discipline ship together. Likewise, compliance is treated as engineering, not marketing.
Have a project
in mind?
Work with the same security software developer who built SafeKey and the API tooling above
Whether you need a security product, API tooling, or a web or desktop app, you can reach out directly. After that, I usually reply the same day.