Operations Systems Architect

I design the systems
that make your tools
talk to each other.

Knowledge bases, PM platforms, CRM architecture, Slack workflows, AI automation - connected into one coherent operating system. Not one perfect app. Clean integrations, real operational improvements.

10+
Years in tech & ops
8x
Quoting speed increase
30%
Automation gains monthly
90%
Company-wide adoption
What I Build
01

Knowledge Systems

Information architecture, SOPs, client playbooks, decision logs, meeting-to-knowledge pipelines. Notion, Confluence, or custom. Governed so it stays accurate and searchable.

02

Operations Platform

PM platform design and implementation. Slack workflows, shared inbox integration, Google Sheets sync. Everything wired so status changes propagate and nothing falls through.

03

AI & Automation Layer

n8n automation pipelines, AI search/Q&A systems, context architecture that makes AI outputs reliable. Browser automation for repetitive flows. The layer that makes everything compound.

Selected Work
Full Operational System - From Zero
Lux Design + Manufacturing | Head of Technology | 2024 - Present

The Problem

A growing FF&E manufacturer with no operational systems. Everything lived in email, scattered spreadsheets, and people's heads. Team across 4 countries with no shared tools, no SOPs, no knowledge base. Every process was manual, every handoff was a risk.

What I Built

  • CRM architecture (HubSpot) with custom data models and pipeline automation
  • Project management system (Jira) with cross-team workflows
  • Structured knowledge base: SOPs, client playbooks, decision logs
  • Slack integration layer for real-time ops capture
  • Google Workspace integration with Drive and Sheets sync
  • n8n automation pipelines connecting all systems
  • AI-assisted estimation engine (FF&E IQ) - reports in minutes, not weeks
30%
Automation increase monthly
6
Team across 4 countries
Min vs Wks
Estimation time
CRM Platform & Integration Architecture
Ergeon | Product Manager | 70+ engineers, ~500 headcount

The Problem

A fast-scaling construction tech company with a CRM held together by workarounds. Manual quoting eating 45+ minutes per lead. Escalations consuming ~50% of possible profit. Disconnected tools, no integration strategy, teams working around the system instead of with it.

What I Built

  • Owned the CRM platform (Pipedrive + custom Django) - data models, API specs, integration architecture
  • Integration portfolio via Zapier and direct APIs between CRM, operational tools, and analytics
  • Automated quoting engine with dozens of composing variables (materials, labor, permits)
  • Automated escalation system with classification, routing, and resolution tracking
  • Jira-based task management designed for company-wide adoption
  • Complete SOP documentation and team training
8x
Quoting speed
25%
Escalation reduction
40%
Lead-to-quote increase
90%
Jira adoption
AI Context & Knowledge Systems
Multiple projects | Architecture + Build

The Principle

AI search and generation are only as good as your information architecture. I design the context layer - constrained inputs, clean metadata, governed source documents - that makes AI outputs reliable. Not magic. Structure.

What I've Built

  • AI Designer: natural language prompts generating production templates from constrained component libraries (React Email, Storybook)
  • AI-first Django workflow: descriptions to code changes with validation pipeline
  • Context architecture for AI estimation engine - structured inputs producing reliable outputs
  • AI-assisted knowledge capture: meeting transcripts to structured KB entries
  • NotebookLM and Claude-based search systems over operational documents
30%
More automated hours
75%
Loading time decrease
Automated Escalation System
Ergeon | Product Manager | Construction operations at scale

The Problem

Escalations were eating roughly 50% of possible profit. No classification system, no routing logic, no visibility into patterns. Issues bounced between teams, resolution times ballooned, and the same problems kept recurring because nobody was tracking root causes.

What I Built

  • Escalation classification system (type, severity, root cause tagging)
  • Automated routing to the right team based on classification
  • Resolution tracking with SLA timers and escalation paths
  • Pattern analysis dashboard surfacing recurring issues
  • Feedback loop: top escalation causes fed back into process improvements and SOPs
  • Training program for frontline teams on new workflow
25%
Escalation reduction
6 to 4 wks
Project turnaround
Millions
In saved profit
Company-Wide Task Management System
Ergeon | Product Manager | Sales, Construction, HR, Admin Ops

The Problem

Every department ran on different tools, different processes, different definitions of "done." Sales tracked work in spreadsheets, construction used email threads, HR had their own system. No single source of truth for who was doing what, no cross-department visibility, and no way to automate handoffs between teams.

What I Built

  • Jira-based task management system spanning Sales, Construction, HR, and Admin Ops
  • 100+ different task types mapped to department-specific workflows
  • 5 integrations: Google Calendar, Email, Slack, Pipedrive CRM, and Django backend
  • Automated task creation and status sync across systems
  • Cross-department handoff automation (e.g., sale closed in Pipedrive triggers construction tasks in Jira)
  • Department-specific dashboards with shared visibility layer for leadership
90%
Company adoption
100+
Task types
5
System integrations
20%
Change order reduction
Satellite-Based Quoting System & Draft Editor
Ergeon | Product Manager | Rolled out to 200+ people across 2 largest teams

The Problem

Quoting took 2.5 days on average. Reps were manually measuring properties, calculating materials, looking up regional pricing, and assembling quotes by hand. The process was slow, error-prone, and couldn't scale. Two of the company's largest teams (100+ people each) were bottlenecked on this single workflow.

What I Built

  • Satellite-based property measurement integrated into the quoting flow
  • Draft Editor allowing reps to build, adjust, and preview quotes in real time
  • Pricing engine with dozens of composing variables (materials, labor rates by region, permit requirements)
  • Automated calculation pipeline replacing manual spreadsheet work
  • Rolled out to 2 largest teams in the company (200+ people total)
  • Mentored 2 junior product managers through the project end-to-end
2.5d to 20m
Avg quoting time
40%
Lead-to-quote increase
15%
Lead-to-win increase
200+
Users across 2 teams
Remote Design Operations at Scale
Microsoft | Design Project Manager | 1000+ projects delivered

The Context

Managed a fully remote design team of ~15 people across time zones - one of the earliest remote management roles in a predominantly in-office Microsoft culture. Responsible for delivery pipeline, sprint coordination, and stakeholder communication across multiple product teams.

What I Delivered

  • support.microsoft.com - design delivery across the support platform
  • docs.microsoft.com - technical documentation design system
  • Microsoft Teams - design contributions during early rollout
  • Windows icon redesign - new generation of Windows icons
  • Backlog management, sprint cycles, and cross-team coordination for all workstreams
  • 1000+ projects delivered on time across the portfolio
1000+
Projects delivered
~15
Remote team size
4
Major product lines
Tools & Stack
Django
Postgres
BigQuery
Python
SQL
Docker
Git
n8n
Zapier
Make
Retool
Metabase
Claude AI
NotebookLM
HubSpot
Pipedrive
Jira
monday.com
Notion
Confluence
Slack APIs
Google Sheets
Shopify
Figma
About

Architecture-trained. Tech-wired. I spent nearly a decade working under a co-founder of oDesk/Upwork across three companies. The most important thing he taught me: data, data, and more data points for everything. No gut feelings. No "I think." Measure it, prove it, then decide.

That principle shaped how I work. When I see a broken process, a manual bottleneck, a system held together by workarounds and good intentions, I can't leave it alone. I will map it, instrument it, find where the data says it's failing, and rebuild it.

I sit between engineering and business. I speak both languages fluently. When things are chaotic, I'm the person who brings structure - not by adding meetings or documents, but by building systems that make the chaos legible.

Background

  • MSc Applied AI (in progress)
  • MArch Architecture & Urbanism
  • CS50 AI + Databases - Harvard

Experience

  • Head of Technology - Lux D+M
  • Product Manager - Ergeon
  • Design PM - Microsoft
  • Associate PM - EzHome

Based in

  • Europe (GMT+1/+2)
  • Working globally

Your tools should compound,
not just coexist.

Let's build your operating system.

Get in touch

I prefer working globally via Upwork - where both client and freelancer are protected.