Jesse Lennon

I build systems that build systems.

I'm an AI-augmented engineer. I design multi-agent systems, distributed infrastructure, and software products, then use AI to build and operate them.

My daily workflow runs through a team of specialized AI agents I designed from the ground up. They handle architecture, implementation, DevOps, strategy, content, research, and coordination. I make the decisions, define the architecture, write the test specs, and debug when things break. The agents help me ship at a pace that used to require a full team.

Before this, I spent 8 years in operations and quality assurance, including QA leadership at Coinbase across NFT Marketplace, Institutional Custody, and Prime Brokerage. That foundation taught me how systems fail, which turns out to be the most useful thing you can know when building systems that work.

Currently Building

CRAWL In Development

Autonomous Multi-Agent Platform

Jesse
Architecture, Decisions, Testing, TDD
Scheduler
Agents own cycles via Redis
Event Router
Redis pub/sub messaging
Model Router
Local → Haiku → Sonnet → Opus
Token Extraction
Compress outputs before expensive models
Eval Scoring
3-level: structural, local, cloud
Memory
PostgreSQL + pgvector
Ender
Commander
Jace
Architect
Carl
Builder
Athena
Strategist
Bob
DevOps
Donut
Content
Fitz
Research

7 specialized AI agents built on Anthropic's Claude API. Four-tier model routing (local Ollama at $0 through Opus), token extraction middleware that compresses tool outputs before expensive models, three-level eval scoring on every invocation, and PostgreSQL + pgvector for institutional memory. Governance primitives: peer review gates, scope guards, budget enforcement. Used daily to build everything on this site. Currently being rebuilt as a standalone platform.

  • Claude API
  • Agent SDK
  • Multi-Agent
  • 4-Tier Model Routing
  • Token Extraction
  • pgvector
  • Redis Pub/Sub
  • Governance
  • TDD

WalletGuardAI Researching

Smart Contract Security Auditor

AI-powered smart contract auditor that analyzes Solidity contracts for vulnerabilities, gas optimization, and best practice violations. MVP complete, parked while LLM audit capabilities matured. Currently exploring whether modern models clear the accuracy bar for security-critical analysis.

  • Node.js
  • React
  • Ethers.js
  • LLM Analysis
  • Solidity

Chancy Alpha

Decentralized Lottery Protocol

A transparent, human-only lottery on World Chain where 90% of every pot goes to the winner. World ID proof-of-personhood enforces one ticket per verified human per draw. Witnet VRF provides verifiable on-chain randomness. Gasless entry via ERC-2771 meta-transactions so users never touch gas fees. Automated keeper bot runs daily draws at 00:00 UTC. Built as a World App MiniApp with Next.js 15 frontend.

  • Solidity
  • Next.js 15
  • World Chain
  • Witnet VRF
  • World ID
  • ERC-2771
  • Hardhat
  • MiniKit SDK

Shipped Projects

Canopy Market Intelligence

Canopy Live

Market Intelligence Mini-App

Built from empty repo to live in 5 days. A crypto market intelligence mini-app for World App (30M+ users). AI-generated market briefs, verified-human conviction polling, curated news feed, educational learn system, and gamification through voting badges and accuracy tracking. World Chat integration for viral sharing. 500+ users, 100K+ impressions. Internationalized across 6 languages with on-chain USDC payments.

  • Next.js 15
  • TypeScript
  • Neon Postgres
  • Drizzle ORM
  • Claude Batch API
  • World ID
  • i18n (6 langs)
  • USDC Payments
Silvic Trading Ecosystem

Silvic Paper Trading

Autonomous AI Trading Ecosystem

A 15-container microservice trading platform in paper trading validation. Three algorithmic strategies on fast loops. Three LLM-powered sentinels tune parameters on slower cycles. A capital allocator scores fitness and redistributes toward whatever is working. 1,200+ tests under strict TDD.

  • Python
  • Docker
  • Redis
  • PostgreSQL
  • WebSocket
  • Next.js
  • GitHub Actions
  • Hetzner VPS

Experience

Program Manager, CX Technology and AI

2023 - Present

Tools for Humanity


  • Led the implementation of Zendesk's advanced AI bots for a global support org. Ran a 5% live traffic test, used real user flow data to rebuild the bot from scratch with structured use-case management. Drove containment from ~55% to a peak of 88%, sustaining ~82% average.
  • Own the full Zendesk platform: administration, triggers, automations, 6 AI bots, WFM, 5 QA programs, and cross-platform workflows (macro management, help center badge automation, Slack-to-Zendesk ticketing, Jira administration, Notion integrations).
  • Picked up iOS (Swift) and Android (Kotlin) Zendesk SDK integration with zero allocated engineering resources. Set up build environments from scratch (Bitrise CI/CD, Lokalise i18n, app emulators). Built A/B/C testing infrastructure with PostHog feature flags for safe production experimentation.
  • Built 6 custom Zendesk apps: a fix for a platform-level SDK authentication bug affecting all custom-auth customers, a sanctions compliance app that auto-tags and escalates restricted-country cases, and a dynamic bug tagging system that routes around native platform limitations.
  • Built a provisioning platform managing 800+ BPO agents across Zendesk and Sana LMS. Time-driven and on-edit triggers keep profiles synced: role changes repush groups and skills, status changes auto-provision or deprovision across both platforms with full audit trails. Domain and role-based access whitelisting for roster permissions. Cut provisioning errors by ~95%.
  • Designed a localization pipeline between EduMe and Lokalise, managing 120K+ translation keys across 20+ languages. Built a dynamic workbook for real-time import/export of translations to work around EduMe's rigid ingestion format.

QA Team Lead, Ecosystems

2021 - 2023

Coinbase


  • Led QA across NFT Marketplace launch (6-month project from inception), Institutional Custody, and Prime Brokerage. Built every QA program from scratch: scorecards, scoring rubrics, auditing cadence, performance management, and training. Navigated novel content moderation challenges unique to NFTs.
  • Built and managed the Critical Error Remediation program end to end, including weekly CE Aging reporting to cross-functional stakeholders. Program passed a regulatory audit with zero findings.

Senior Quality Analyst, Client Services

2020 - 2021

Coinbase


  • Designed, launched, and managed QA programs across Institutional Custody (Cold Storage), Exchange, and Institutional Onboarding. All programs built to adhere to KYC, AML, OFAC, and SOC 1 and 2 compliance standards.
  • Redesigned the Institutional Onboarding QA program to report on application-level effort and surface new KPIs, improving visibility into onboarding pipeline quality.

Earlier roles available on  

Skills

AI and LLM

  • Multi-Agent Systems
  • Anthropic SDK
  • Prompt Engineering
  • Tool Use
  • MCP Servers
  • Model Routing

Backend

  • Python
  • Node.js
  • PostgreSQL
  • Redis
  • WebSocket
  • REST APIs
  • Event-Driven
  • Drizzle ORM
  • Neon Postgres
  • Lambda
  • DynamoDB
  • Pytest

Frontend

  • Next.js 15
  • React
  • TypeScript
  • Tailwind CSS
  • Auth.js
  • Recharts
  • Vitest
  • i18n

Infrastructure

  • Docker
  • GitHub Actions
  • Vercel
  • Hetzner VPS
  • Terraform
  • AWS
  • S3 / CloudFront
  • Linux

Web3

  • Solidity
  • Smart Contracts
  • Hardhat
  • OpenZeppelin
  • Ethers.js
  • World Chain
  • World ID

Foundation

  • System Design
  • TDD
  • Distributed Systems
  • Microservices
  • Technical Leadership
  • Program Management
  • Compliance (SOC/KYC/AML)
  • Zendesk Platform

Get in touch

Available for fractional AI engineering engagements, architecture reviews, and consulting.