I’m Noel, a software engineer interested in how the latest AI tooling can be applied to real problems. I like taking new capabilities and stress testing them against actual workflows to see what holds up.
I work at Maven AGI, a Boston based AI startup building enterprise AI agents for customer support. We build enterprise-grade AI agents that unify systems, knowledge, and actions so organizations can automate and improve customer support and experience at scale. My role centers on building the tooling that helps customers design, configure, and ship better agents, faster.
Outside of work, I experiment with new agent architectures, proactive behaviors, evolving capabilities, and new Claude Code skills. I explore how agents can integrate with legacy systems and extend themselves within defined boundaries.
I’m also building Quotient, a receipt scanning bill splitting app. It gives me a full stack environment to apply the same ideas end to end, from messy inputs to reliable outputs.
Most of my work sits at the intersection of AI capability and practical execution.
Built an AI-powered receipt scanning and bill-splitting app with automated Venmo links and SMS reminders.
Created multiple web apps to help the Kansas Department of Transportation implement their Road Usage Charge Pilot.
Recreated a product using an LED Matrix and a Raspberry Pi Pico.
Built a multi-agent AI planner that breaks down high-level goals into actionable calendar events.
Created an iOS application for the University of Vermont Medical Center to educate fall-prone individuals.
Created a platform that allowed renters to review their landlords and apartments to increase transparency in the rental market.
Redesigned and implemented an SEO-optimized website, resulting in an estimated 300% increase in page-visits.
Optimized a simulated robot for locomotion in diverse environments using evolutionary algorithms.
Developed a simple compiler enabling translation of high-level programming language into x86 assembly language
Using Arduinos, sensors, and pumps connected via Bluetooth, created a device to keep plants watered and fertilized.
Collaborated with a small team to recreate the game Plants vs. Zombies
Developed a fully-functional version of the WWII German encyphering machine Enigma in Python.
© Noel Desmarais
February 2026