Scott Guthart
Software Engineer — Full-Stack, AI/LLMs, Data Science
Full-stack software engineer with a backend focus, independently architecting and self-hosting AI-driven applications end to end — from Postgres data models to production infrastructure — for Fortune 100 companies including CVS, Johnson & Johnson, Procter & Gamble, and L'Oréal.
Experience
AI Full Stack Engineer — Edelman
- Independently architect, build, and self-host full-stack AI applications end to end — React, Next.js, Tailwind, and shadcn on a Supabase (PostgreSQL) backend, deployed on AWS EC2 — for client teams serving Fortune 100 companies such as Samsung, Johnson & Johnson, and CVS.
- Built an automated media-monitoring pipeline that scrapes top-tier outlets like The New York Times and The Wall Street Journal and uses LLMs with structured output to classify coverage and model topics, with human-in-the-loop review so strategists can validate and refine results.
- Partner directly with internal teams to gather feedback and iterate quickly, developing AI-natively with tools like Cursor and Claude to ship and refine features rapidly.
Lead Software Engineer — Radius
- Founded and led the company's first software engineering team, hiring, mentoring, and establishing the engineering practices and architecture used to scale solutions across the organization.
- Built a full-stack application (Python, React) integrating OpenAI's LLM to automate coding of thousands of open-ended survey responses, with analysts reviewing outputs in a human-in-the-loop workflow.
- Shipped a statistical-toolkit app deployed on Linux via a REST API, improving a five-person analytics team's efficiency by 20%.
Full Stack Engineer — Radius
- Built and deployed a market-simulation app — still in use by clients such as CVS, Johnson & Johnson, Procter & Gamble, and L'Oréal — on a Linux cloud server.
- Replaced an Excel-based system with a custom project-management app for 200+ projects, featuring Salesforce integration, automated reporting, and change tracking (MySQL backend).
Statistical Analyst — Radius
- Automated multivariate statistical techniques in Python (NumPy, scikit-learn), improving analyst efficiency by 15%.
Selected project
MovieTable
Public full-stack movie discovery app combining critic and user ratings with AI-powered recommendations and weighted scoring. It is available to explore publicly and includes a recommendation system designed to showcase the full-stack and AI work behind it.
Stack: Next.js 16, React 19, TypeScript, Tailwind CSS 4, shadcn/ui, ReUI, TanStack Table, Supabase/PostgreSQL; self-hosted on a DigitalOcean Ubuntu VM with Coolify (PaaS) and Cloudflare DNS
MovieTable was built with an AI-native workflow using v0, Claude Code, and Vercel.
Earlier — Clinical & Research
- Physical Therapist, Per Diem — John T. Mather Memorial Hospital, Port Jefferson, NY
- June 2017 – August 2018
- Physical Therapy Aide — John T. Mather Memorial Hospital, Port Jefferson, NY
- December 2013 – June 2017
- Research and Wellness Intern — RRAMP Lab, Stony Brook, NY · Healthcare research and community education
- January – March 2017
- Clinical Intern — Keep Moving Forward Inc., Garden City, NY · Outpatient pediatric neurology
- June – August 2016
- Claims Manager / Adjuster — Tokio Marine America · Medical and disability workers' compensation case management
- 2011 – 2012
Education
Publications
- Augmented mindfulness intervention for overweight / obese children and adolescents
- Clustering of Clusters in Ensemble Segmentation: Which Technique Works Best?
- Feasibility of an Interdisciplinary Family-Centered Adolescent/Young Adult Obesity Intervention
- Novel Biopsychosocial Osteoarthritis Intervention: Case Study