HackerRank’s AI Day 2025: Product Launch Recap

AI is changing what it means to be a software developer—to be an orchestrator of AI agents. AI also impacts trust in the hiring process — did the candidate follow the instructions? This means the skills you need to build and the way you hire and upskill needs to change.

AI is making what was once impossible, now possible. It was not possible to test people’s thought process at scale. It was not possible to have a proctor in every remote room. It was not possible to create an entire hiring event in minutes. But now, with AI, it is.

At AI Day 2025, we unveiled a set of features that will enable you to hire and upskill the next generation of developers. Here’s a sneak peek into what was showcased:

Proctor mode

Proctor mode is like having a human proctor for take-home assessments. It sets the test instructions, detects suspicious activity in real time, and submits a comprehensive report that includes a full-session replay.

• Real-time detection: If suspicious activity is detected, the Proctor immediately notifies the candidate and requests an explanation
• Session replay: To differentiate between honest mistakes and misconduct, a full session replay, including code playback and image capture, is available in one place
• Transparent reports: When a session is flagged, Proctor mode provides a detailed explanation for review

AI interviewer

An AI designed to conduct first-round interviews that go beyond simply ensuring code correctness by evaluating depth of knowledge, problem-solving approach, code quality, and ability to work with AI. Equipped with contextual hints and progressively challenging questions, it closely simulates a real interview experience.

• Gives hints when you’re stuck: Just like a human interviewer, the AI interviewer gives real-time hints when a candidate is stuck. It guides candidates without revealing answers.
• Ramps up complexity: The AI builds on the original context with progressively harder questions as the candidate provides correct answers.
• Understands thought process: The AI doesn’t just grade answers, it is sophisticated enough to judge thought process by asking follow-up questions to see how candidates think.

We initially launched the AI Interviewer for our developer community so that developers could prepare for interviews and receive feedback. We have conducted over 12,000 interviews, and our reviews have been overwhelmingly positive. Now, you can use this for the initial screen, which gives you deeper signals and an even stronger candidate experience. 

Engage assistant

Use the new Engage assistant to rediscover past candidates with the skills you’re looking for and source new candidates from our database. Build a marketing campaign in minutes to connect authentically with developers.

• Rediscover past candidates: The Engage Assistant reviews your past candidates and matches them to the skills you’re currently hiring for.
  
• Automated marketing campaigns: The assistant creates targeted campaigns that include an on-brand microsite, a curated list of rediscovered candidates, and a developer-friendly email campaign.
  
• Publish to the HackerRank Developer Community: You can now publish your campaign to the HackerRank candidate database, which includes the members of our 26M+ developer community

AI tutor

SkillUp’s AI tutor helps upskill your developers through real-world tasks, creates structured learning paths, and certifies them to be next-gen developers. It supports multiple learning paths across foundational, intermediate, and advanced levels.

• Teaches without giving the answer: Developers can tackle real-world practice challenges with AI-driven support that offers hints and explanations without giving away answers
• Assesses real world skills: The learning path is constructed with real world tasks on code repos to resemble tasks completed in the actual job
• Certifications to show off your skills: The certifications are also based on real world problems thus giving you confidence in their ability to do great work

ASTRA benchmark

We’ve built the HR-ASTRA benchmark to assess LLMs on real-world problems, measuring correctness, consistency, efficiency, cost & communication. Our evolving leaderboard offers continuous comparisons and updates.

• Real-world datasets: The evaluation focuses on the model’s ability to perform complex tasks across the software development life cycle.
• Primary focused on correctness and consistency: To assess the production reliability of the model, we prioritize median standard deviation with k=32, rather than relying on the industry-standard pass@k.
• Dive deeper into skill leaderboards: To assist in model selection for various tasks, performance is reported at both the skill level and on a global leaderboard.

All of the above will be rolled out gradually starting in April 2025. 

If you’re interested in being at the forefront of next-generation hiring, contact your Customer Account Manager for early access to our latest innovations.

You can also access the session recording of HackerRank AI Day here.