The turning point in Cathy Lai’s coding-interview practice was not solving more difficult problems faster. It was slowing down: clarify the question, trace an example, explain the logic, and only then write code. Her account offers a practical routine for candidates who feel tempted to jump straight into hard LeetCode problems, though it is one person’s experience—not a proven formula for interview success.
Start with a problem you can learn from
Lai says she began with easy, AI-generated exercises and raised the difficulty gradually rather than starting with problems that damaged her confidence. She aimed for two to three problems a day, depending on difficulty. That number describes her personal routine, not a recommended quota or a measure of readiness.
As an Amazon Associate I earn from qualifying purchases.
The useful principle is to choose a challenge that lets you practice the full reasoning process. If a problem feels opaque, stepping down in difficulty can give you room to build the habits that carry over to harder questions.
Recommended Free Tools
Make your reasoning visible before coding
For each problem, Lai’s approach moves from understanding to a traceable plan, then to implementation. The sequence makes it easier to spot a mistaken assumption before it becomes a tangle of code.
#1 Best Overall
- Careercup, Easy To Read
- Condition : Good
- Compact for travelling
Clarify the assumptions
Write down what the prompt does and does not specify. Ask questions about input, expected output, edge cases, and constraints rather than silently choosing an interpretation. In an interview, state those assumptions aloud so the interviewer can correct them.
Check the coding setup
Before tackling the main logic, verify that the environment behaves as expected with a small dummy function and a test output. This separates setup or syntax trouble from the problem-solving work.
Trace an example by hand
Walk through a sample input one step at a time. Track the values that change on each iteration, and note which variables are needed to represent the state. Lai puts it this way: “Trace the algorithm manually: Walk through the example input step-by-step to identify every variable needed across iterations.”
Name the uncertainty
If you are stuck, say what you are trying to determine: whether the algorithm needs a flag, a running total, a value for each group, or some other state. Naming the uncertainty gives you a concrete question to resolve instead of leaving the problem as a vague block.
Check the proposed logic against the trace
Run the example through your plan and verify when each value is set, reset, or accumulated. If the trace contradicts the logic, revise the plan before implementing it. Lai’s implementation advice is: “Only write code once the logic is proven—this prevents getting bogged down in syntax while still problem-solving.”
Implement in small steps and debug calmly
Once the logic is clear, translate it into code. Pseudocode and explicit state tracking can help keep the plan visible while you implement. Test incrementally instead of waiting until the whole solution is written: simple print debugging can reveal the contents of a data structure or show where a value first goes wrong.
Unexpected output is a normal debugging signal, not proof that the entire attempt has failed. Compare the observed values with your hand trace, locate the first divergence, and check the relevant update or condition. This turns debugging into a continuation of the reasoning process.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsPractice the interview conversation as well as the solution
Lai recorded some practice sessions and reviewed her pacing, explanations, and overall presence. Recording can help you notice habits that are hard to catch while solving, but her account does not establish that recording improves interview outcomes. A practice partner can provide another perspective; a commenter on the post recommends working with someone experienced in hiring, including on behavioral interviews.
Best Value
Another commenter describes solving Codewars challenges and then reading and explaining other people’s solutions aloud. That is a discussion suggestion, rather than a step in Lai’s original routine. Either kind of practice can help you rehearse describing a solution clearly, but the post does not compare these approaches or show that one is more effective.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Use AI as a practice aid, not an authority
In a reply, Lai describes organizing questions in a project, starting a separate conversation for each problem, pasting her solution into ChatGPT for critique, and stating the difficulty she wanted. This is her reported workflow; it does not show that an AI tool will reliably judge difficulty or give accurate instruction in every case.
If you use an AI assistant, treat its feedback as something to verify. Check suggested logic against the prompt and sample cases, and make sure you can explain the reasoning without relying on the assistant’s answer.
Free tools Windows power users keep installed
One-click scans. No signup required.
What this approach does—and does not—establish
Cathy Lai’s September 16, 2026 DEV Community post is a short personal account, labeled AI-assisted. It describes a way she prepared, not a controlled study. It reports no measured improvement, interview pass rate, or comparison of AI practice with human coaching. The method is best understood as a concrete routine for making problem-solving, communication, and debugging more deliberate—not a guarantee of a job offer.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
