Tosh Meston / Staff Software Engineer · AI Product Engineering
I bring decades of mobile, web, and product engineering to human–AI interaction. From Microsoft and Myspace to Taskrabbit, I turn complex systems into software people can use.
Education / In progress
M.S. Computer Science · Artificial Intelligence specialization
Individual project, CS6750
Human-Computer Interaction
01 / Experience
Large product teams. Developer platforms. Production apps. I bring technical depth, product judgment, and hands-on delivery to the same table.
Software Design Engineer
9 years
Built features for Outlook Web Access (OWA), bringing email to the browser for millions of users. Part of Microsoft’s first cohort of Web Design Engineers (WDEs), with work spanning NetDocs, Outlook, and usable, performant web interfaces across Office and Exchange.
Web applications · Outlook + Exchange · Product engineering
Staff Software Engineer
Current role
Own production mobile features for the Dolly consumer and Helpers apps in Swift and Kotlin. Partner with product and design, guide architecture, and improve performance and reliability.
iOS · Android · Mobile architecture
Software Architect / Development Manager
5 years across two stints
Led Myspace’s Platform API engineering team, delivering OpenSocial APIs, third-party application integrations, and features for major partners. Later architected OAuth 2.0 authentication and mobile push notifications for the Myspace app.
OpenSocial · Developer APIs · OAuth 2.0
02 / Human–AI interaction
What does the model know? What is it doing? How can a person steer it? I explore those questions by building the interfaces and the systems underneath.
I designed and built the agent execution loop, tool orchestration, model-provider integrations, and streaming interface. Coding App connects file tools, an integrated browser, and local or hosted models in a workspace where people can inspect the agent's context and actions.
Explore browser use and agent feedbackFrom prototype to understanding
01Make the system's facts and guesses visible, editable, and deletable.
02Turn model actions and structured results into understandable UI.
My human–computer interaction work at Georgia Tech informs these prototypes: what do people need to understand an AI system, correct it, and stay in control?
03 / Selected client work
Custom coaching tools for the Dallas Mavericks. An official Seattle Sounders FC fan app across seven seasons. Software built for real teams, real workflows, and sustained use.
I turned InfiniteHoops Whiteboard into a custom coaching platform for the Dallas Mavericks staff.
The work expanded a simple drawing tool into a private iPad workflow for storing plays, building multi-page sequences, replaying ideas, and distributing the app to Mavericks coaching staff iPads through Apple Business Manager.
Design and development by me
I worked on the official Seattle Sounders FC app across multiple seasons, from 2016 through 2022.
The app brought matchday information, news, video, rosters, standings, and voting into one supporter experience.
Mobile development by me. Design by Elizabeth Rhodes and Seattle Sounders FC.
04 / Studio + independent work
Since founding Broadside Digital in 2014, I've led a distributed team and delivered web and mobile products for clients, alongside products of my own.
Client product · Mobile + services
Built native iOS and Android apps and supporting web services for personalized bonsai learning. Design by Bonsai Mirai.
View the iOS appFounder · Design + engineering
Created a pickup basketball network and a suite of coaching apps for drawing plays, planning practice, and tracking games.
Explore InfiniteHoopsClient product · Native iOS
Built a focused calendar app with everyday scheduling and event editing in a deliberately reduced interface. Design by Rationale.
View Minimal CalendarA deeper foundation
Swift, SwiftUI, Kotlin, React, Next.js, TypeScript, Node.js, Python, and AWS. From interaction details to APIs and architecture.
Georgia Tech · CS6750 Human-Computer Interaction. Individual project selected as “Exemplary,” April 2026.
MIT xPRO · Designing and Building AI Products and Services. Exemplary Assignment, May 2025.
Additional coursework in tree-based learning, including decision trees and bagged random trees; reinforcement learning with Q-learning; and algorithmic fairness, including bias evaluation and mitigation.
M.S. Computer Science, Artificial Intelligence specialization (in progress)
Georgia Institute of Technology
B.S. Physics, Mathematics minor
Brigham Young University
CITI Program · Human Research, Social / Behavioral Research. January 2026.