HackerRank just dumped its hiring agent on GitHub (interviewstreet/hiring-agent), and within a day it pulled nearly 3,000 stars and a 789-point Hacker News thread. Then a developer actually ran it — and the thing fell apart in public.
What it is and why it broke
It’s an AI resume screener. Feed it a PDF, it parses the text, calls an LLM six times to pull out skills, projects, experience, and awards, then spits out a score. The catch: run the identical resume repeatedly and the score swings between 66 and 99. Even at temperature 0 it stays non-deterministic. As the tester put it — if a company draws the line at 85, he gets filtered out 65% of the time. Same resume. Pure dice roll.
The real lesson
The damage isn’t random. Technical skills score consistently because it’s a checklist — you know React or you don’t. Projects and open source flail, because that’s a subjective call, and 65% of the total weight sits there. The experience prompt is literally two lines, no rubric. So a senior engineer with no public GitHub gets buried under someone with two internships and a side repo.
This is the uncomfortable part nobody wants on the box: LLMs are bad at fine-grained subjective judgment, and bolting one onto hiring turns a filter into a lottery.
You Might Also Like
- Insforge Hits 1 on Product Hunt and 3600 Github Stars is This What Agent Native Backends Look Like
- Openviking Treats ai Agent Memory Like a File System and 9k Github Stars say its Working
- Tradingagents 33k Stars for an Open Source Framework That Simulates an Entire Trading Firm With llm Agents
- From Claude Flow to Ruflo 22k Stars 5900 Commits and the Multi Agent Swarm Taking Over Claude Code
- Oh my Claudecode Turns Claude Code Into a 32 Agent dev Team and 11 8k Stars Agree
Top AI Product 
Leave a comment