Reliable, efficient systems — APIs, auth, Postgres, and deploys that don't surprise you in production.
I'm Samuel Mount, a Rutgers CS student and full-stack engineer. I like measurable wins: faster CI, fewer outages, and schemas that stay honest as the product grows.
- 5person team; I led milestones alpha → v1.0
- −48%CI build time (Docker layer caching)
- 3releases, zero unplanned downtime
- 49IT staff across 11 lab locations
- 1,100+SN tickets resolved (VPN, Azure, MFA, M365, …)
- 80+consultants trained in RCA & customer service
Supervisor of the Year — OIT Busch & Cook/Douglass Labs, 2025.
I do my best work owning the backend end-to-end — schema, APIs, auth, deployment, and the guardrails that keep releases clean. At Change Now I was the sole backend owner for a 5-person team: zero unplanned downtime across three releases. At OIT I kept 49 staff and 11 labs running under real operational load — that's where I learned that fixing root causes beats fixing symptoms.
Most of my work is backend infrastructure, not consumer chat. Real exceptions: integrating an external AI TA over APIs and WebSockets (CollabLab × Agoro), and designing OpenAPI/Swagger contracts on Change Now so the system stays machine-readable and automation-ready from day one.
- Deploys you can repeat: CI gates, Docker layers, prod parity.
- Auth that fits the client: JWT + HTTP-only cookies where it matters.
- Postgres integrity: constraints and tests that catch drift early.
Projects
A mobile app that tracks workouts and turns logs into clear, visual progress.
- Cut CI build time 48% with Docker layer caching
- Zero unplanned downtime across 3 releases (domain + Nginx + HTTPS on EC2)
- CI gates: Playwright + Jest required before merge
- JWT + HTTP-only cookie dual-mode auth to support both API clients and browser-safe sessions
- OpenAPI/Swagger so contracts stay explicit — easier for teammates and for future AI/automation clients
- Postgres schema designed for referential integrity across users, exercises, and workout logs
A Java desktop app to manage a university vehicle fleet: reservations, trip history, and department cost reports.
- MVC separation kept business rules testable and the UI straightforward to iterate on
- Black-box test strategy across UI and model logic with JUnit
- Singleton pattern for shared state where it simplified coordination without duplicating logic
- Inheritance + polymorphism to reduce repeated business rules in the model layer
Partner integration that exposes CollabLab teaching surfaces (whiteboard, codeboard, audio routing) to an external AI TA, with CollabLab isolated from Agoro-specific orchestration logic.
- Clear contract boundary: TA automation plugs in without coupling CollabLab internals to Agoro
- Single integration path for multi-surface control (board + audio) instead of one-off hooks
- Keep Agoro/orchestration outside CollabLab — expose stable capabilities over sockets/API so either side can evolve independently
- Favor explicit control messages over implicit UI hacks so routing and permissions stay auditable
A lightweight dashboard that pulls live weather API data and compares AQI metrics across multiple cities.
- Dynamic reflow UI: 8 widgets reposition automatically as cities are added
- Color-mapped AQI (green → crimson) to make severity obvious at a glance
- UI layout prioritized clarity and quick comparisons over dense charts
- Kept dependencies minimal to ensure fast startup and easy sharing