遇见数据集

The 25 hardest companies to interview for in 2026: difficulty, offer rate and experience data

收藏
Zenodo2026-09-28 更新2026-10-01 收录
官方服务:

资源简介:

This dataset lists the 25 companies with the highest candidate-reported interview difficulty in Dataford's interview data, as of a Sept. 4, 2026 snapshot. For each company it gives the difficulty score, the number of rated reports, the offer rate, the share of positive interview experiences, median tech pay where published, and the employee work-life rating. The figures are taken from Dataford's ranking of the 25 hardest companies to interview for in 2026, published Sept. 3, 2026. Summary McKinsey has the highest difficulty score, 3.78 on a five-point scale, across 1,000 rated reports. Juspay (3.68) and Jane Street (3.54) are second and third. The median company in the wider data scores 3.0, and every company in this file scores 3.33 or higher. Consulting and quant trading firms account for 14 of the 25 rows. Google, at 3.35 and 19th, is the highest-ranked Big Tech company, and Meta is 25th at 3.33. Amazon (3.21), Apple (3.16) and Microsoft (3.13) are below the cutoff. Offer rates range from 6 percent at Hudson River Trading to 48 percent at A.T. Kearney. Positive-experience rates range from 12 percent at Canonical to 72 percent at Bain. Meta has the highest median pay in the file, $360,236. Columns rank; company; industry; difficulty (mean candidate-reported score, 1 = very easy, 5 = very difficult); rated_reports (number of reports with a difficulty rating, at least 300 per company); offer_rate_pct (share of fully parsed reports with an outcome that ended in an offer); positive_experience_pct (share of reports marked as a positive experience); median_pay_usd (company median across tech-role pay bands, blank where the company does not publish bands); work_life_rating (employee rating from 1 to 5, blank where the source gives none). Method Interview reports are candidate-submitted interview experiences collected in Dataford's pipeline from sources including Glassdoor, Blind and LeetCode. Near-duplicate company names were merged and review-farm entries removed. Employee ratings use a company's aggregate Glassdoor profile where it has 500 or more reviews, and otherwise Dataford's own review sample of 100 or more reviews. Pay uses leveled total-compensation ladders where published and otherwise the best-supported posted ranges with 15 or more data points. Posted ranges understate equity-heavy packages. All reviews are volunteered, so the sample is self-selected. Compiled by the team behind Dataford.

提供机构:
Zenodo
创建时间:
2026-09-28
二维码
社区交流群
二维码
科研交流群
商业服务