My company recently opened positions for entry-level developers and testers on LinkedIn. We offered competitive packages up to 20 lakhs annually. The response was massive - over 12k applications came in.
We had to screen out most applicants upfront because their profiles didn’t match our requirements. We didn’t want to waste everyone’s time with unnecessary interview rounds.
During technical rounds, we tested fundamental programming concepts and data structures like arrays, stacks, graphs, searching algorithms, etc. We even let people use ChatGPT to write solutions. But here’s the problem - when we asked candidates to explain their code or discuss performance metrics, most couldn’t answer.
It seems like many applicants just copy AI-generated code without actually learning how it works. They can’t walk through their own solutions or optimize them.
Are we being too strict with our evaluation criteria? Or do junior developers really need to focus more on understanding concepts instead of relying on AI tools? Looking for advice on improving our hiring approach.
Your screening process might be filtering out candidates who could actually succeed in the role. While conceptual understanding is important, many junior developers learn best through hands-on experience rather than verbal explanations under pressure. Consider restructuring your evaluation to include practical debugging exercises or collaborative problem-solving sessions where candidates can demonstrate their thought process naturally. The fact that you received 12,000 applications suggests strong market interest, but zero hires from 450+ interviews indicates a potential mismatch between your assessment methods and real-world job requirements. Perhaps evaluate candidates on their ability to research, adapt, and learn from mistakes rather than perfect explanations of algorithms they may rarely use in actual development work.
honestly sounds like your expectations might be too high for entry level? i mean if someone can use chatgpt effectively and solve the problem thats still a skill right? maybe try asking them to debug broken code instead - that shows real understanding better than explaining algorithms imo
wow 450+ interviews with zero hires sounds brutal! im curious tho - did you try pairing candidates with your devs for like 30 mins instead of just asking them to explain? maybe they understand better when actually coding together? also what happend when you gave them simpler problems first?