Evan Langlais

Park Predictor is an end-to-end data platform that continuously collects, processes, and visualizes ride wait times across 50+ US theme parks — from Disney and Universal to Cedar Fair and Six Flags. A Go-based data collector polls live wait times every 5 minutes from parks nationwide, storing raw time-series data in PostgreSQL and computing derived analytics through a multi-layer processing pipeline.

The processing pipeline goes beyond raw wait times, computing real-time velocity and trend detection (is this line getting longer or shorter?), park-level crowd pressure scoring, pairwise attraction correlation analysis (which rides fill up together?), and crowd redistribution impact measurement (when a ride goes down, where do the crowds go?). These derived metrics power a Nuxt 4 analytics dashboard where users can monitor live conditions, explore historical trends, and compare wait patterns across rides on interactive time-series charts.

The long-term vision for Park Predictor is to leverage this growing dataset and analytics pipeline to build a predictive model — ultimately delivered via mobile app — that can tell park visitors when ride wait times are likely to drop, helping them make smarter decisions about when to queue and when to wait it out.

• Frameworks: Nuxt 4 (Vue 3), Nitro
• Languages: Go, TypeScript
• Data: PostgreSQL, Supabase, Unovis
• Infrastructure: Azure Static Web Apps, Docker

• Visit parkpredictor.com

Family Archives is a self-hosted family media platform built to give my family a private, centralized place to upload, organize, and relive photos and videos from family events. Members authenticate via passwordless email login, browse events in a clean dashboard, and view media through a built-in theater with full video playback and photo viewer — all without handing our memories over to a third-party service.

Under the hood, the platform is powered by a Nuxt 4 frontend deployed as an Azure Static Web App and a dedicated Go media processor running on my home NAS. Uploaded files are chunked and sent to a self-hosted Garage S3 instance, where the Go processor picks them up, transcodes video via FFmpeg, optimizes images, generates thumbnails, and archives originals — all automatically. Secure communication between the frontend and the Garage S3 endpoints is handled through a Cloudflare Tunnel, keeping everything accessible without exposing my home network.

While currently built for my family, the platform is designed with extensibility in mind — the group-based access control, event organization model, and self-contained processing pipeline are architected to potentially serve other families as a product in the future.

• Frameworks: Nuxt 4 (Vue 3), Nitro
• Languages: Go, TypeScript
• Data: PostgreSQL (Supabase), Garage S3
• Infrastructure: Azure Static Web Apps, Cloudflare Tunnel, FFmpeg, Docker

Software Engineer II -- Azure Core Networking, Virtual Networking (VNET) team

Core engineer on Azure Accelerated Connections (NSDI ‘23) from inception through production GA — a greenfield platform that disaggregates stateful network functions off individual hosts into shared pools of programmable hardware. Designed and delivered key platform subsystems, and became a primary technical resource for cross-org design coordination, customer escalations, and enabling feature teams to build on the platform.

Designed and delivered a high-throughput billing pipeline processing millions of dollars in daily revenue, requiring novel approaches to line-rate traffic classification under hardware constraints and end-to-end integration with upstream billing systems.

Architected and shipped a new resource provisioning mode (now publicly documented as Azure Virtual Network Routing Appliance) that unlocked critical AI customer workloads by adapting an established programming pipeline to support an entirely new usage pattern under tight timelines.

Drove platform production readiness by defining end-to-end reliability KPIs, building observability dashboards and regression testing infrastructure, and leading systematic investigations to measurable improvement. Served as a primary escalation point for cross-product investigations across the platform suite.

Software Engineer III -- Productivity Suite Team

Built and shipped Office product extensions on the Productivity Suite team, delivering features that augmented customer workflows to improve efficiency and effectiveness directly within Excel, Word, and PowerPoint.

Partnered with product managers to define requirements and deliver end-to-end solutions using C# within the Razor WPF ASP.NET framework, integrating with Office via COM and VB.NET.

Freelance Software Engineer

Collaborated directly with CEO on product vision, viability, design, and implementation across the full stack to close competitive gaps and launch new revenue-generating features.

Owned product design and implementation across backend, frontend, and infrastructure, delivering features including vendor price-comparison tools and expanded product offering customization.

Abstract

Co-authored paper presenting Sirius, a system for disaggregating stateful network functions (firewalls, NAT, load balancing) into shared pools of P4-programmable NIC appliances, achieving 5-10x improvement in connections-per-second for Azure VMs. Now in production as Azure Accelerated Connections.

Presentation

• NSDI Page
• PDF
• Slides

Led a team of 6 to build an IoT platform for a Connecticut generator company, allowing technicians to remotely monitor and control 7,000+ generators via a web dashboard. Designed a five-stage data pipeline spanning Raspberry Pi (Go/MODBUS), a Go gateway server, Kafka/Spark processing, OpenTSDB storage, and an Angular frontend.

• Technical Overview

A Vue 3 / Ionic mobile app with a Node.js backend (hosted in Azure) that helped participants track fundraising progress and event information for UConn’s dance marathon.

A hobby game built during college using Phaser.io and Cordova. Players match colors on screen using the Delta-E 2000 algorithm. Still available on the App Store.

• Download in App Store