← back

Selected work.

Production systems, shipped for paying customers — in AI, multi-tenant SaaS, and the recovery of deployments that went sideways. A fuller account than the homepage.

01

FXVPS

Founder & Lead Engineer · 2017 — present

Trading-infrastructure SaaS, run end to end — now an AI agent and reliability platform.

Context

A multi-tenant VPS and trading-tooling SaaS for retail algorithmic traders, operated solo across two brands — product, infrastructure, billing, customer comms, incident response, and the AI tooling layer. Underneath it all sits a self-healing provisioning platform spanning multiple hypervisors and regions; the last year has been agent and reliability engineering on a live customer base where downtime means a customer’s automated strategy stops trading.

Outcome

Phantom data and accounting drift eliminated, alerting moved from none to four real-time channels, provisioning and billing failures designed out at the platform layer, and the codebase de-risked from per-portal forks to shared modules.

What I built
  • Multi-channel Trade Journal alert engine — Telegram, ntfy, browser SSE, and HMAC-signed webhooks off one multi-tenant Go + PHP stack. The same release also hardened multi-tenant data isolation.
  • An LLM content-audit pipeline that read 126 customer-facing emails, flagged compliance and consistency issues, and shipped corrections to 38 active customers with zero failures.
  • A multi-platform trade reconciliation system that caught $4k of accounting drift and $540k of phantom data across 7 accounts through cross-platform duplicate detection, then re-anchored portfolio P&L to the broker as source of truth.
  • A self-healing provisioning pipeline — golden templates, IP-annotation reconciliation, and fail-closed account creation with a logon self-test — that retired a recurring class of duplicate-allocation and silent-password-failure bugs.
  • A billing-lifecycle layer that auto-cancels stale invoices and reconciles against the host daily, with dunning rules tuned so live trading accounts are never suspended prematurely.
  • A headless MetaTrader compile-and-publish pipeline for the customer journaling indicators, plus a brand-neutral Trade Journal hook that replaced per-portal patches with one shared module.
  • Go
  • PHP
  • MT4/MT5
  • WHMCS
  • Anthropic Claude
  • DeepSeek
  • MCP
  • Multi-region provisioning
02

Farm to You

Engineering Lead (contract) · 2024 — present

Producer-direct marketplace SaaS — feature-architecture leadership across a nine-person team.

Context

Engineering lead across web, API, and delivery layers for a producer-direct marketplace, on contract via UK Ltd. Year one stabilised a fragile platform; year two shifted to feature-architecture leadership and opening new revenue lanes — wholesale, partner-led growth, and producer self-service — across a nine-person remote team.

Outcome

Major features and a new wholesale lane shipped behind flags with controlled rollout; 985 issues migrated from Jira to Plane; a Tuesday-to-Tuesday sprint cadence aligned to live market operations.

What I built
  • The v4 seller and delivery chain — producer onboarding, order flow, and delivery management — rolled out behind progressive feature flags so a live marketplace kept trading through the migration.
  • The Market Partner Program: multi-tenant schema, partner-introduced producer prefill, full-funnel attribution, and an admin approval queue, piloting across three markets.
  • A two-sided wholesale channel — Farm Direct Sourcing — where producers list inventory and respond to buyer sourcing requests, opening a B2B lane alongside the consumer storefront.
  • An SEO foundations rebuild — sitemap, structured data, canonical normalisation, page-speed — and a /sell rework that replaced placeholder platform statistics with real producer proof.
  • PHP
  • JavaScript
  • MySQL
  • Feature flags
  • Multi-tenant SaaS
  • Playwright
  • Plane
03

WeSync

Co-Founder & CTO · 2021 — present

Independent SaaS and AI implementation-recovery service.

Context

WeSync audits failed or stalled enterprise software and AI deployments, identifies root cause, and delivers fixed-fee remediation without depending on the original vendor — Big-4 finance rigour paired with Fortune-500 technical architecture.

Outcome

A productised, vendor-independent recovery offering — co-founded with a former Deloitte consultant carrying $150–340M programme experience.

What I built
  • A 21-day Implementation Recovery Audit: a fixed-scope technical and process autopsy delivered as a report.
  • A remediation engagement model for independent execution of the recovery plan, plus an ongoing fractional CFO + tech-architect retainer.
  • Architecture
  • Technical due diligence
  • Fixed-fee delivery
04

Company.com

Chief Technology Officer · 2013 — 2015

A $32M turnaround that became Inc 5000 #5.

Context

CTO of a bundled operational-cloud-software company for SMBs, with partners including OfficeMax and Radio Shack. Inherited a costly J2EE stack and a hybrid onshore/offshore team.

Outcome

The work carried the company to #1 fastest-growing in Georgia and #5 on the Inc 5000 in 2016 — fifth-fastest-growing private company in America.

What I built
  • Migrated the web properties from J2EE and PostgreSQL to PHP (Yii) and MySQL, and rebuilt 10 business products.
  • Cut annual staffing cost from $1.8M to $650k and reduced hardware infrastructure cost tenfold, managing a hybrid team of 13.
  • PHP / Yii
  • MySQL
  • Cloud infrastructure
  • Team leadership
// before that

Twenty years, condensed.

// let's talk

Have a system worth getting right?

main@eaymon.com

or find me on LinkedIn