TaskFlow — Debugging Technical Interview

Welcome! This is a timeboxed (~60 minute) technical interview built around a small task-management library called TaskFlow. There are two identical implementations in the same repo — Python and Java — so pick whichever language you're most comfortable in.

The interview has two parts:

1. Debugging (~35 min) — The library ships with a failing test suite. A few bugs have been planted. Your job is to find and fix them until the tests are green.
2. Feature (~20 min) — Once tests pass, you'll add a new feature. This part is open-ended: use any external resources you like (docs, Google, StackOverflow, AI assistants such as Copilot/ChatGPT/Claude, etc.). We care about how you approach the problem, not whether you memorized an API.

We're not looking for perfection. We want to see how you read unfamiliar code, form hypotheses, verify them, and communicate as you go. Think out loud.

---

What is TaskFlow?

A tiny in-memory task tracker. Tasks have an id, title, priority (1 = low, 2 = medium, 3 = high), a list of tags, and a completed flag. The TaskManager class lets you add tasks, look them up, complete them, remove them, count them, filter by tag, list pending tasks, and sort by priority.

The two languege implementations behave identically — same classes, same methods, same bugs.

---

Setup

Clone the repo and pick a language.

Prerequisites

Language	Needs
Python	Python 3.9+ and pytest (pip install pytest)
Java	JDK 17+ and Maven 3.8+

Python

cd python
python -m pip install pytest        # once
python -m pytest -v                 # run the tests

Java

cd java
mvn test                            # compiles and runs the tests

Tip: run a single test while iterating.

• Python: python -m pytest tests/test_manager.py::test_complete_task_marks_completed -v
• Java: mvn -Dtest=TaskManagerTest#completeTaskMarksCompleted test

---

Part 1 — Debugging (~35 min)

1. Run the test suite. You should see multiple failures.
2. Read the failing tests in tests/ (Python) or src/test/ (Java) to understand the intended behavior.
3. Open the source (taskmanager/manager.py or src/main/java/com/example/taskmanager/) and fix the bugs.
4. Re-run until everything is green.

There are six planted bugs: four are easy to spot from a single failing test (a loud exception or an obviously-wrong result), and two are subtler — they only surface on an edge case, so the failing test won't point straight at the buggy line. The test file groups them into "easier" and "harder" sections to help you pace yourself. Fix the source, not the tests — the tests describe correct behavior.

As you work, tell us: what does the failing test expect, what did you observe, what's your hypothesis, and how did the fix confirm it?

---

Part 2 — Add a Feature (~20 min)

Once the suite is green, pick one feature below (or propose your own) and implement it, including at least one test. This half is intentionally open — reach for whatever tools and references you'd normally use.

Suggested features (in rough order of scope):

• Due dates & overdue list. Add an optional due date to tasks and a get_overdue(today) / getOverdue(today) method returning uncompleted tasks past due.
• Keyword search. Add search(keyword) that returns tasks whose title contains the keyword, case-insensitively.
• Tag summary. Add tag_counts() / tagCounts() returning a map of tag → number of tasks with that tag.
• Priority bump. Add bump_priority(id) / bumpPriority(id) that raises a task's priority by one level (capped at high), and decide what should happen at the cap.

Walk us through your design choices, edge cases, and how you'd extend it further with more time.

---

What we're evaluating

• Debugging method — reading code, isolating faults, verifying fixes.
• Communication — narrating your reasoning and trade-offs.
• Code quality — clean, readable changes that match the surrounding style.
• Feature judgment — sensible design, edge-case awareness, and a test that proves it.

Good luck — and remember to think out loud!