Skip to content
armanet.xyz
armanet.xyz

I build things, then I run them.

Nearly five years across blockchain consulting in Paris, QA engineering in Singapore and applied AI. Now I build LLM products solo, and run them on hardware I maintain myself, with agents doing the operations behind my review.

The through-line is not a stack. It is finishing: getting something past review, past first launch, past the week where it breaks, and then keeping it alive long enough to learn what it actually costs.

SingaporeEnglish (fluent) · French (native)Open to roles

Career

  1. 2026 to now

    Independent

    AI systems and infrastructure

    Running real products without a team

  2. 2024 to 2026

    iLex

    QA Engineer

    Fintech · interbank loan management · 15-person team

  3. 2022 to 2023

    BearingPoint

    Blockchain IT Consultant

    Blockchain task force for public institutions and private clients

  4. 2020 to 2021

    Plebicom

    Software Engineering Intern

    Paris

  5. 2017 to 2022

    EPITA

    MSc Computer Engineering

    France

Skills

AI systems
Agent orchestration, tool and permission design, LLM evaluation (labelled reference sets, Cohen's kappa, recall), model cascades, RAG, MCP, Anthropic API, local LLMs (Ollama, Qwen)
Languages
TypeScript, Python, Swift, Solidity, C/C++, Java
Product stack
React, React Native, Expo, SwiftUI, Next.js, Node.js, Express, Supabase, PostgreSQL, SQLite, MinIO
Infra & QA
Proxmox, Docker, Traefik, Cloudflare Tunnel, Tailscale, ZFS, GitHub Actions, Karate, API test automation, fuzz testing
Blockchain
Solidity (EVM, ERC721/1155/4337), Solana (web3.js, Anchor, Jito), XMTP
Soft skills
Teaching and training (700+ people trained), client delivery to public institutions, end-to-end ownership without a team, written operating procedures

No proficiency bars. A percentage next to a language name is a number somebody made up about themselves. The projects are the evidence.