Selected Work
A few examples of the kinds of systems Belden Studio builds.
This is a concise sample, not an exhaustive portfolio. The common thread is practical software for teams that need working products, reliable systems, and strong technical judgment.
AI agent product
AI Outreach Agent
Built core systems for an AI outreach product that indexed multiple social graphs and orchestrated campaigns across LinkedIn, email, phone, and SMS.
- Owned the product experience from onboarding through campaign execution.
- Built an agent-driven campaign builder combining conversational and form-based workflows, with live UI updates as campaigns were defined.
- Embedded AI throughout the interface to analyze context, auto-fill inputs, generate targets, and guide campaign configuration.
- Developed a WebSocket-powered dashboard showing campaign state, recipient activity, and system feedback in real time.
- Implemented GPT Realtime voice agents for structured campaign setup and automated candidate screening.
Full-stack operations
Operations Automation Platform
Led stakeholder interviews, defined technical strategy, and built a full-stack operations system supporting more than $2M in monthly transactions across hundreds of B2B clients.
- Served as the sole owner across engineering, product, and DevOps.
- Built a Next.js application with internal inventory tooling, customer-facing ordering surfaces, role-based authentication, and automated data pipelines.
- Architected a real-time sales work order system connecting inventory, sales floor workflows, warehouse fulfillment, and QuickBooks invoicing.
- Implemented Telegram and WhatsApp workflows for customer communication and sales notifications.
Real-time media systems
Live Streaming Platform
Worked on high-performance web clients for WebRTC broadcasting and live streaming at multi-million user scale.
- Led cross-functional development of realtime web clients using React, Redux, and GraphQL.
- Designed and deployed a client-side ML pipeline using TensorFlow for real-time video quality optimization, executing millions of in-browser inferences at scale.
- Owned web client security, driving vulnerability identification and remediation across client and API surfaces.
Enterprise video product
Async Video Communication Platform
Led development of an enterprise platform for asynchronous video communication, including a proprietary multi-layer video player and core systems for publishing, playback, and content management.
- Built end-to-end product features including metadata and privacy controls, commenting, and video discovery.
- Integrated React-based publishing workflows into a macOS desktop application via embedded webviews, enabling seamless cross-platform video creation.
- Established QA processes from scratch, onboarding and mentoring the first QA engineer.
Production UI systems
Consumer Security Product
Built core production UI components powering key pages and user flows for a widely-used consumer security product, used by millions of users.
- Translated Figma designs into high-fidelity, performant, and accessible experiences.
- Partnered closely with design and product to implement pixel-accurate layouts, animations, and responsive behavior.
- Designed and implemented CMS-backed component systems, enabling dynamic content updates and reducing reliance on engineering for day-to-day site changes.
Semantic search / AI
AI-Powered Developer Search
Built an AI-powered search engine enabling engineers to retrieve technical results using semantic language.
- Built the frontend in React and Redux, partnering with UI/UX designers to craft intuitive search interfaces.
- Collaborated with backend engineers on a Python/Flask and Elasticsearch RESTful API to improve search speed and accuracy.
Computer vision / AI
Computer Vision Web Application
Integrated computer vision models into a video-based web application, bringing real-time visual analysis directly into the browser.
- Built React frontend video components that surfaced computer vision model outputs in real time.
- Integrated TensorFlow models with GraphQL and AWS cloud infrastructure.
Independent product
CookingBuddy.AI
Built a voice-first cooking assistant to explore hands-free conversational UX, realtime AI interaction, and product behavior in a live user-facing context.
- Used as a practical vehicle for refining voice interaction patterns and realtime assistant behavior.
- Focused on interaction quality, pacing, and usability rather than novelty alone.
- Helpful as a proof point for voice product thinking and execution.
Something here relevant to a project you're working on?
The fastest way to figure out fit is a short conversation.