Clarence Elliott

Clarence Elliott

I help engineering teams of 10 to 40 turn AI coding tools into shipped throughput instead of unused licenses.

Fractional CTO. Ten years of production engineering, five at Google, and in Claude Code every day.

Your team has Copilot, or Cursor, or Claude Code. Adoption looked good for about a month. Delivery looks roughly the same as it did before. That is the problem I work on.

Scope a diagnostic See the work
Clarence Elliott
10
Years shipping production software
5
Years at Google
10,000+
Daily users of the AI feedback service I led backend for
Daily
In Claude Code, building Project Plus

It is usually not the tools. Writing code got faster, review stayed exactly as slow, and nobody re-measured the constraint.

Consulting

Start with a two-week diagnostic.

I spend two weeks inside how your team actually works with AI coding tools, and you leave with a clear picture of where delivery is being lost and what to do about it, in order. Ongoing work runs as a monthly retainer, and most of it starts here.

Scope a diagnostic

The diagnostic

  • What I look at. Who actually uses the tools and who quietly stopped. Where throughput is lost between writing and merging. What the codebase does and does not give the tools to work with, and where that has already produced something risky.
  • What you get. A written assessment and a ranked roadmap: what to fix first, what it costs, and what it returns.
  • Terms. Fixed scope. Fixed price, from $5,000. Two weeks. No open-ended discovery.

Work History

Where I’ve built things.

  1. 2025 — Present

    Project Plus Co-Founder & CTO

    An AI-native real estate platform that helps experienced agents automate property analysis and reclaim their time.

    • Lead all technical architecture, AI system design, and product engineering across the founding team.
    • Built the platform’s flagship capability — a computer-vision workflow that reads listing images to identify property features and generate an automated quality rating, turning a subjective manual assessment into an instant score.
    • Translate insights from 20+ discovery interviews with practicing agents directly into product requirements and shipped features.
    • Own cloud architecture, data modeling, and automation pipelines end to end.
    • Built and shipped with an AI-assisted workflow (Claude Code) from day one, which is where the diagnostic comes from.
    • Next.js
    • Supabase
    • Vercel
    • n8n
    • Claude Code
    • Computer Vision
  2. 2020 — 2025 Remote

    Google Software Engineer

    • Led backend for the AI feedback rating-and-storage service inside Google’s internal Ads customer-support CRM, used daily by 10,000+ support agents.
    • Product owner for the service, translating the needs of internal support staff into technical requirements and roadmap decisions.
    • Built internal access and identity tooling for the Ads organization, using FlumeJava pipelines to analyze logs and report on data-access provisioning.
    • Authored microservice architecture and release-plan practices adopted by three sibling teams.
    • Mentored six engineers on CI/CD pipeline architecture, cutting average service development time by a full quarter.
    • Java
    • FlumeJava
    • Microservices
    • CI/CD
  3. 2018 — 2020 Chicago, IL

    AllCampus Backend Developer

    Higher-education marketing and enrollment-services provider.

    • Built greenfield features on a custom CRM, improving internal efficiency by optimizing external API usage and automating report generation.
    • Built a tool that parsed institutional performance data and auto-generated tailored market-research decks, streamlining Sales-team workflows.
    • Developed and managed asynchronous task queues with Celery and RabbitMQ.
    • Python
    • Celery
    • RabbitMQ
    • AWS
Show 2 earlier roles 2015 — 2018
  1. 2017 — 2018 Chicago, IL

    TekSystems Java | Python Engineer

    Supported financial-services clients, modernizing risk-reporting processes to keep Bank of America compliant with federal mandates.

    • Implemented a Python data-sourcing library that centralized application access to uploaded spreadsheets and data from Teradata, DB2, and Oracle.
    • Migrated seven critical spreadsheets onto the library with minimal user impact.
    • Built a payment website in Angular and TypeScript with the RESTful services behind it.
    • Python
    • Java
    • Angular
    • Teradata
    • Oracle
  2. 2015 — 2017 Chicago, IL

    JP Morgan Chase Technology Analyst — Software Developer

    Global Customer Information Service team.

    • Reengineered legacy COBOL services into Java microservices delivering PII/SPII data to internal applications across the bank.
    • Validated new services against legacy logic to ensure logical equivalence and correct database interaction.
    • Configured Nagios network monitoring for 100+ production web services.
    • Java
    • COBOL
    • SoapUI
    • Nagios

Education

  • B.S., Computer Science University of Illinois Urbana-Champaign