Tech Interviews Got Harder in 2026: AI Is Up 4x, and LeetCode Never Left

Companies didn't swap LeetCode for AI. They added AI on top. The share of technical interview reports mentioning AI, LLMs or generative AI rose more than 4x in 18 months, from 1.6% in Q1 2025 to 7.2% in Q3 2026. Over the same period, LeetCode-style questions held steady at about 30% of technical reports.

That comes from 59,738 technical interview reports at a fixed panel of 639 companies collected by Interview Query. The new material is a set of rounds that barely existed two years ago: AI interviewers (up 14x since 2024), AI-assisted coding rounds, and LLM case questions. For ML engineers, AI came up in nearly half of 2026 interviews. For data scientists, 15%.

1. AI mentions in technical interviews are up 4x in 18 months, and about 20x since 2023. In Q1 2023, 0.4% of technical interview reports mentioned AI. By Q1 2025 it was 1.6%. By Q3 2026 it was 7.2%. Most of that rise happened in the last four quarters.

2. The same companies tripled their AI mentions. To rule out "we just collected different companies," we compared the same 533 companies in 2024 and 2026. Across all roles at those companies, the share of interview reports mentioning AI went from 0.9% to 3.0%. Among the 267 companies with at least 30 reports in both years, the AI share rose at 73% and fell at only 12%.

3. LeetCode did not go away. LeetCode or algorithm questions appeared in 30.2% of technical interview reports in Q1 2025 and 30.5% in Q3 2026. Companies kept the coding screen and added the AI round on top of it.

4. AI interviewers grew 14x. Reports describing an AI interviewer, an AI recruiter, or an automated AI screening call went from 0.5 per 1,000 interview reports in Q1 2024 to 7.0 per 1,000 in Q3 2026. Candidates describe a bot that asks follow-up questions based on their answers, usually as the first step before a human ever gets on a call.

5. AI-assisted coding rounds went from almost zero to 1 in 90 technical interviews. In Q1 2025, 2 of 5,025 technical reports described a coding round where you work with an AI assistant. In Q3 2026 it was 37 of 3,368, or 1.1%. These go by names like "AI coding," "AI-assisted pair programming," or "agent design," and you are graded on how you direct the tool and check its output as well as on whether the code runs.

For role-level numbers we used only first-hand reports submitted directly to Interview Query in 2026, where the candidate picked their own role (1,412 reports). Here is the share that mentioned AI, LLMs or GenAI:

• ML or AI engineer: 47% (n=81)
• Data scientist: 15% (n=352)
• Software engineer: 9% (n=681)
• Data engineer: 8% (n=89)
• Data or product analyst: 6% (n=209)

Data scientists hit AI in their interviews about 2.5 times as often as data analysts. In practice that looks like a recruiter screen asking about your GenAI experience, a RAG or LLM evaluation question in the ML round, or a case that asks how you would measure an AI feature. Analyst loops are still mostly SQL, metrics and product sense.

Data science loops are also lighter on algorithms than engineering ones: 19% of data scientist reports mentioned LeetCode-style questions, versus 52% for software engineers. So for data scientists, the new AI material is a bigger share of what changed.

A few companies stand out. Among companies with 100+ reports in 2026, AI came up in 21% of Ramp reports (30 of 144, versus 1 of 171 in 2024), 15% at Intuit (27 of 177), and about 10% at Snowflake and Databricks. At Meta, it was 5.7% of 578 reports, up from about 1% in 2024. These are small counts per company, so treat them as examples, not a ranking.

1. Keep practicing the fundamentals. SQL, stats and coding are still most of the loop.
2. Prepare one strong AI story. Be ready to explain a project where you used or evaluated an LLM: what you measured, what failed, and what you would change.
3. Practice coding with an assistant. If you get an AI-assisted round, the skill being tested is breaking the problem down, prompting clearly, and catching wrong output.
4. Treat the AI screen like a real interview. It is often the first filter. Answer in full, structured sentences, because the follow-up questions are built from what you say.

• Trend data: 189,483 public interview reviews from candidates at a fixed panel of 639 companies, dated January 2023 to September 2026, collected and classified by Interview Query. We use the same company panel for every quarter so changes are not driven by which companies were collected.
• Technical reports (59,738) are reviews that mention coding, SQL, Python, algorithms, system design, machine learning, statistics, or a technical round.
• AI mention means the review text includes AI, LLM, GenAI, generative AI, GPT, ChatGPT, Copilot, Claude, OpenAI, or prompt engineering. Random spot checks found about 9 in 10 matches were about AI in the interview itself. The rest were incidental, such as a company with "AI" in its name.
• Role data: 1,412 first-hand reports submitted to Interview Query between January and October 2026, grouped by the role the candidate selected.
• Limits: A mention is not the same as a tested skill. Candidates who describe an AI round without using the word "AI" are not counted, so the round-type numbers are floors. Counts for individual companies are small.

How common are AI questions in data science interviews in 2026? In first-hand reports submitted to Interview Query in 2026, 15% of data scientist interviews mentioned AI, LLMs or generative AI (54 of 352 reports). That compares with 6% for data and product analysts and 47% for ML and AI engineers.

Are companies still asking LeetCode questions in 2026? Yes. LeetCode or algorithm questions appeared in 30.5% of technical interview reports in Q3 2026, about the same as the 30.2% in Q1 2025. AI rounds were added alongside coding screens.

What is an AI interviewer and how common is it? An AI interviewer is an automated screen where a bot asks questions, often by voice or video, and generates follow-ups from your answers. Reports describing one rose from 0.5 per 1,000 interview reports in Q1 2024 to 7.0 per 1,000 in Q3 2026.

Can you use AI tools like ChatGPT or Copilot during a coding interview? Sometimes, and it is becoming more common. 1.1% of technical interview reports in Q3 2026 described an AI-assisted coding round where using an assistant is part of the test, up from 0.04% in Q1 2025. Most coding rounds still do not allow it, so ask the recruiter before the interview.

How much has AI changed technical interviews since 2023? The share of technical interview reports mentioning AI rose from 0.4% in Q1 2023 to 7.2% in Q3 2026, roughly 20x, across the same 639 companies.