It seems that many processes, from interviews to real work, are increasingly manipulated. I've noticed a pattern with candidates employed by certain consulting companies, especially in Texas and New Jersey. These companies often recruit low-cost labor from India, craft fake resumes, and submit them to platforms like LinkedIn.
During interviews, candidates use tools like HDMI dual-screen setups, ChatGPT, Otter AI, or Fathom AI to cheat and secure jobs. These consulting firms even fabricate green card verifications and other documents, enabling them to crack most interviews unless the candidate is exceptionally unskilled.
Once hired, these companies often delegate the actual work to individuals in India, paying them as little as $500 while profiting $4,000–$5,000 per month from the arrangement.
We uncovered this issue when we began conducting on-site interviews. While these candidates can handle medium-level LeetCode problems during virtual evaluations, they struggle with basic tasks, like implementing a LinkedList or solving simple LeetCode problems, in person.
Alarmingly, these consulting companies are becoming more sophisticated over time. This raises a critical question: how can genuinely experienced candidates compete in such a landscape?
I keep holding out hope that one day my totally genuine, slightly rusty, slightly nervous, takes all 40 minutes to solve the Leetcode medium style will be seen as so refreshing and honest I’ll be an insta-hire.
They are taking advantage of the incompetence at the workplace you're at. That's just what business is and has always been. If you're a fool, you'll be separated from your money.
Unfortunately so many people lie about experience that you need to so some sort of whiteboard test just to see if the candidate really is fluent in the language they are claiming 5 years experience with. It can be a really simple test.
In my two decades of experience, I've never seen another software engineer implement a linked list or even use a linked list. There are better, and more interesting, questions to be asking.
I personally wouldn't expect someone to implement one (end cases easy to mess up if they are stressed), but writing a function to reverse one (foreach, pop front, push front) is enough to catch the liars. You can argue about how often a std::list vs std::vector is a performance win, but I'd run a mile from any developer who wasn't highly familiar with the basic data structures provided by any language they are claiming to be fluent in.
The only real requirements to "never use a linked list" are a) use a language where some kind of contiguous-storage-based sequence (array, vector, whatever you want to call it; Python calls it a list, even) is built in (or in the standard library); plus b) not ever need to remove O(1) values from the middle of a sequence in O(1) time while preserving order.
But arguably, a candidate who hasn't ever had to contemplate the concept of "linked list" but can derive the necessary ideas on the spot given the basic design, has some useful talents.
I've done this. It can be hit or miss. Get a great team with a strong lead and you'll love them. Unfortunately there's quite a bit of opportunity over there so once you've trained them up, they're always looking for their next (better paying) gig with their new skills. It's rare if folks last past a year on your team.
There are so many incredibly talented software engineers in India that want to stay in India for family/cultural reasons. The best setups I have seen have one very reliable senior person who experience working in EU/NA, then returned home. They can help with the cultural barriers with more junior hires. Further, if you pay 20% more than your competition, you can get way better candidates. My experience is also pretty similar with offshore teams in China, but their English skills are worse (on average).
During interviews, candidates use tools like HDMI dual-screen setups, ChatGPT, Otter AI, or Fathom AI to cheat and secure jobs. These consulting firms even fabricate green card verifications and other documents, enabling them to crack most interviews unless the candidate is exceptionally unskilled.
Once hired, these companies often delegate the actual work to individuals in India, paying them as little as $500 while profiting $4,000–$5,000 per month from the arrangement.
We uncovered this issue when we began conducting on-site interviews. While these candidates can handle medium-level LeetCode problems during virtual evaluations, they struggle with basic tasks, like implementing a LinkedList or solving simple LeetCode problems, in person.
Alarmingly, these consulting companies are becoming more sophisticated over time. This raises a critical question: how can genuinely experienced candidates compete in such a landscape?