Skip to content

Assignments

Why these assignments exist

Joining a research lab in 2026 is not just "read some papers and start coding." You need to plug into shared infrastructure (OSC compute, W&B tracking, GitHub), adopt the lab's tool stack (polars, altair, uv, PyTorch), and build habits for working with AI coding assistants. These assignments get you there in ~4 weeks.

By the end of onboarding you will have:

  • A published personal academic website that doubles as a portfolio
  • An EDA blog post demonstrating real polars + altair fluency
  • OSC, W&B, and GitHub Education accounts fully set up
  • A neural network you built from scratch — no frameworks
  • A working mental model of package management, dependency resolution, and AI-augmented development

How assignments work

  1. Read the full assignment before starting — each one lists prerequisites, estimated time, and deliverables.
  2. Work through the tasks in order — later tasks build on earlier ones.
  3. Ask for help early — if you're stuck for more than 15 minutes on a setup issue, post in the lab Slack/Teams channel. Setup problems are normal and everyone encounters them.
  4. Publish your deliverables — every assignment ends with a blog post. See the Publishing Guide for the workflow.

Current assignments

  • Assignment 1: Personal Website

    Git, GitHub, VS Code, Quarto, GitHub Pages. Ship a site that's the foundation for every future blog post.

    3–4 hrs · Beginner

    Start Assignment 1

  • Assignment 2a: Exploratory Data Analysis

    polars, altair, scikit-learn. Explore a real automotive intrusion detection dataset and publish the analysis.

    6–8 hrs · Beginner

    Start Assignment 2a

  • Assignment 2b: Research Infrastructure

    OSC, W&B, GitHub Education. ML pipeline mental model. SQL as the foundation of experiment tracking.

    3–4 hrs · Beginner

    Start Assignment 2b

  • Assignment 2c: AI-Augmented Development

    Copilot, LLM interaction framework, CLAUDE.md. Build habits for working with AI assistants effectively.

    3–4 hrs · Beginner

    Start Assignment 2c

  • Assignment 3a: Package Management Concepts

    pip vs conda vs uv. Virtual environments, dependency resolution, lock files. Why this matters on OSC.

    3–4 hrs · Beginner

    Start Assignment 3a

  • Assignment 3b: Neural Network from Scratch

    numpy only. Implement forward pass, backprop, and a training loop. Understand what PyTorch does under the hood.

    5–7 hrs · Beginner–Intermediate

    Start Assignment 3b

Scheduling

  • 2a + 2b + 2c are assigned together — 2 weeks total. 2a is hands-on coding; 2b is lab infrastructure + pipeline concepts; 2c is AI tooling. Part 4 of 2c configures the repo you built in 2a.
  • 3a + 3b are assigned together — 2 weeks total. 3a is conceptual; 3b is hands-on. Both end in a blog post.

Cross-cutting references

  • Publishing Guide

    The single reference for turning an assignment into a blog post: YAML headers, categories, preview/commit/push workflow, portfolio habits.

    Open the guide

  • Student Site Showcase

    Example personal sites built by lab members. Use them for inspiration when designing your own.

    Browse showcase

  • Typst CV Guide (optional)

    Upgrade your CV to a professionally typeset PDF with Typst. Not required — but recommended before job applications.

    Read the guide