Reliability monitoring, wallet operations tooling, and AI-assisted automation for crypto infrastructure. Currently running ops for a 2,000-unit ASIC fleet at ~98% uptime across multiple sites.

Background runs in an unusual order: eight years as a futures broker (NFA Series 3, expired) → 2.5 years in enterprise endpoint security at ESET → mining infrastructure operations. Crypto-native since 2018; building AI agent systems on VPS since 2024.

Languages: Python, Bash, PowerShell. SQL at working level.

Infrastructure: Linux (Debian / Arch / RHEL), Windows Server, Docker, systemd, networking (TCP/IP, DNS, DHCP, firewalls), Tailscale, WireGuard.

Observability: Prometheus, Grafana, structured logging, metric-driven alerting, incident escalation patterns.

APIs: Postman, cURL, REST/JSON, OAuth/JWT, webhook signing, HTTP debugging.

Web3: Bitcoin, EVM, and Solana —> node ops, multi-signature wallet operations, onchain investigation with Etherscan, Solscan, Mempool, Tenderly, and Blockstream.

AI / agents: OpenRouter, LiteLLM proxy, multi-agent orchestration on VPS, Docker backend, prompt engineering, RAG fundamentals.

Support tooling: Zendesk, Jira, Bomgar, NetSuite.

Multi-chain wallet monitoring with sanctioned-address screening and structured alert envelopes designed to feed ticketing, SIEM, or LLM-assisted incident summaries.

Tech: Python, PyYAML, Etherscan API, Blockstream API.

Built for: Detecting large outflows from monitored addresses; flagging interactions with sanctioned wallets and known mixer contracts; producing JSON alert records that downstream systems can consume without further parsing.

What I learned: Bitcoin's UTXO model and EVM's account model need genuinely different "what just happened" logic — you can't fake one with the other. Structuring alert events for consumers I don't control (some hypothetical SIEM, some hypothetical LLM summarizer) was harder than the detection itself; the schema is the contract.

Production-style observability stack for Bitcoin and Monero node uptime, peer connectivity, and infrastructure health.

Tech: Prometheus, Grafana, Docker, systemd, Linux.

Built for: Replacing manual SSH-and-check workflows with dashboards and metric-based alerting. Real-time visibility into node operations with escalation-grade data, not just "is this thing reachable."

Daily-driver scripts for ASIC fleet monitoring: miner health, pool connectivity, thermal status.

Tech: Python, Bash, JSON-RPC.

Built for: Standardizing fleet checks across thousands of units where 5% offline at any given moment is the steady state, not an emergency. Structured output feeds log aggregation; degraded-miner detection runs against historical baselines instead of arbitrary thresholds.

What I learned: "Is this thing okay?" logic at fleet scale has to handle missing data as a normal case, not an error case. Otherwise your alerts page you for the network, not the miners.

Marketplace escrow MVP on Base Sepolia. Explores Chainlink CRE for dispute resolution events and transparent settlement logic.

Tech: Solidity, Python, Chainlink CRE, Base Sepolia.

Status: Work in progress — current focus has shifted to other projects.

Field-ready CLI / DevOps / Web3 reference. Personal notes that became shareable.

Most of my active work lives in private repositories — AI agent infrastructure on VPS (multi-agent orchestration with shared memory, tool routing, and persistent execution), prediction-market trading automation, and a SaaS product in the operations-tooling space. The public projects above are extracted utilities and reference work from this broader stack.

cryptoaijedi@proton.me

USA · US Eastern · Open to global remote and relocation