Jason Azizibuilds AI‑run operations

An AI agent-operations company that designs and deploys teams of AI agents for businesses, with approval gates the owner controls action by action.

• Operating org: leadership and manager agents with defined roles and reporting lines direct parallel execution agents, each working from its own skill set.
• Shared memory: structured hand-offs, decisions and client notes land in one knowledge base the whole team reads, with the people, companies, projects and ideas in it linked to each other, so context carries from one agent, and one day, to the next.
• Mission Control: a status dashboard with a Waiting-on-you queue (approve / deny / answer) and a switchboard that sets each action's gate (allow, ask or block) for the whole team or one agent, behind token authentication, output redaction and private-network access.
• Governance: a 46-skill library (sales, delivery, finance and document-generation tiers) that encodes permission policies and approval gates. Every install starts gated; the owner opens the send, spend and publish gates as each loop proves itself.
• Integrations: custom connectors into the systems a business already runs, semantic search across the team's memory, and an assistant with email ingest and voice-memo transcription.

• Multi-agent orchestration
• Claude Code
• MCP
• Agent SDK
• TypeScript
• Python

Ran a B2B office-technology supply company: sales strategy and forecasting, pricing, national accounts, purchasing, inventory and finance.

• Grew revenue 5x over 2009–2022, from $1.5M to $7.5M, by building an automated office-technology business with nationwide accounts.
• Won back Public Storage (2,800 locations nationwide) and ran all of its operations: order fulfillment, customer support, reporting and receivables.
• Launched an ecommerce store that increased revenue by $2M with no additional sales staffing, and built the Amazon company store from $0 to $2.5M+ in 2 years (150+ items).
• Reached 95% customer retention while cutting customers' toner costs 30% through automated replenishment (HP, Lexmark, Brother and Xerox programs).
• Over 2009–2022, built and led a team of up to 14 direct reports, managed the operating budget and P&L performance, and built a 3-technician repair division that generated $400K+ in revenue and grew warranty revenue 40%.

Full-cycle enterprise sales of hardware, cybersecurity, document management, cloud, backfile conversion, IT and automation services, from first presentation through quarterly business reviews.

• Achieved 140% of sales quota in Q1 2024 and 195% in 1H 2024.
• Increased the territory's software solutions mix from 15% to 30% in 2024.
• Led Los Angeles in net-new logo sales for 2025 and reached annual quota by Q2 2025.
• Owned territory sales forecasting and built Salesforce reports.

A delta-neutral funding-rate carry system in Python, extended to hedged on-chain liquidity, with layered risk controls.

• Scans five perpetual-futures venues (Hyperliquid, dYdX v4, Aster, Paradex, Lighter) with net-of-cost rate math, a history-calibrated backtester and a capacity guardrail.
• Places paired cross-venue orders through signed venue clients (Hyperliquid SDK, dYdX v4 client, Aster API).
• Risk controls: kill switch, circuit breakers, per-venue caps, a dry-run mode, idempotent closes and an hourly book-vs-venue reconciliation guard.
• On-chain module: delta-neutral concentrated-liquidity positions on Uniswap v3, Aerodrome Slipstream and PancakeSwap v3, hedged with Hyperliquid perpetuals, with automated open, close, re-range and gauge staking through signed EVM transactions.

• Python
• Hyperliquid SDK
• dYdX v4
• Uniswap v3
• Aerodrome
• EVM signing
• JSON-RPC