real_work

Tired of grinding random, abstract LeetCode problems? Try doing actual work, for real.

Comment

real_work

Solve problems where they actually happen.

Inspiration

Engineering is more than just passing a coding interview. In the real world, you aren't asked to invert a binary tree on a whiteboard; you're asked to debug a production race condition, optimize a slow database query, or fix a broken CI/CD pipeline. We were tired of the "LeetCode vacuum" where problems are disconnected from actual infrastructure. We wanted a platform where the environment is as much a part of the challenge as the code itself. real_work is our solution.

What it does

real_work provides a marketplace of engineering "labs" that run inside fully isolated Docker containers. Unlike traditional platforms that only give you a text editor, real_work gives you a full Linux environment.

  • Hands-on Labs: Solve challenges ranging from backend performance tuning, distributed systems debugging, dev-ops, web dev, game dev, or penetration testing.
  • Persistence: Progress is saved automatically. When you leave a workspace, your exact container state is committed to an image and stored on a container registry on the cloud, allowing you to resume exactly where you left off later.
  • Entirely Web Based: No need to install anything. All containers are deployed on the cloud and are displayed to the user through a virtual desktop on the site.
  • Performance Benchmarking/Grader: Implement your own custom grading system that connects to our API to handle grading your solution.
  • Community-Driven: Anyone can publish a challenge by linking a GitHub repository or creating a portable Docker container. Users can rate, comment, and discuss strategies for every lab.

How we built it

The platform is a modern full-stack application built with:

  • Frontend: Next.js 15 (App Router) with TailwindCSS for a sleek, engineering-focused UI.
  • Backend/Auth: Supabase (PostgreSQL) for real-time database management, user authentication, and storage (thumbnails/avatars).
  • Infrastructure: A custom Node.js/Docker orchestration layer that manages container lifecycles, port mapping, and state persistence.
  • Challenge Creator: A multi-step wizard that parses GitHub repositories, extracts function signatures, and allows authors to bake custom configurations into reusable Docker images.

Challenges we ran into

  • State Persistence: Moving beyond simple volume mounts to a robust docker commit workflow was difficult. We had to ensure that user progress was captured across sessions without bloating the database.
  • Windows Compatibility: Navigating the differences between Docker on Windows (PowerShell/WSL) and Linux required careful path sanitization and permission handling (chmod in-container vs. Windows paths).
  • Security: Running arbitrary user code is dangerous. We implemented multi-layered isolation to ensure that containers remain sandboxed and system-level access is restricted.

Accomplishments that we're proud of

  • The "Resume" Experience: Achieving a "VM-like" feel where you can close your browser, turn off your computer, and come back to the same terminal state.
  • The Creator Wizard: Building a tool that can take a raw GitHub URL and transform it into a ready-to-solve engineering challenge in under a minute.
  • Unified Thematic UI: Creating a cohesive design language that feels professional and "alive," using consistent animations and interactive feedback. The solving process takes place entirely within the website with no external dependencies, which is crucial for user immersion.

What we learned

  • Docker Internals: We gained deep knowledge of the Docker Engine API, specifically image layering, committing container diffs, and dynamic port allocation.
  • Next.js Server Actions: Leveraging Server Actions for complex infrastructure tasks (like launching containers) allowed us to keep the client light while maintaining a high level of security.
  • Context is King: We learned that learners engage more deeply when they are "solving a fire" in a realistic environment than when they are solving abstract mathematical puzzles.

What's next for real_work

  • Multi-Container Labs: Supporting complex environments like "Debug this Kubernetes Cluster" or "Fix this Microservices Mesh."
  • Leaderboards & Competitions: Real-time engineering sprints where the community competes to solve production incidents the fastest.
  • Company Tracks: Helping companies build internal training labs that mirror their actual tech stacks.

Built With

Share this project:

Updates