About the role

Hiring is one of the most consequential decisions a company makes. 3,000+ enterprises rely on HackerRank to get it right. We are now reinventing how that works for the agentic era. The ML systems that power this platform are not auxiliary features. They are the product.

Open Problems

The agentic era is reshaping every layer of the hiring stack. These are some of the core problems you'll be working across, none of them fully solved.

• Chakra: Building an autonomous AI interviewer that conducts, adapts to, and evaluates technical interviews end to end.
• Integrity: Detecting fraud and suspicious behavior across multiple signal types. The ways candidates game assessments change frequently, and the models need to keep up.
• Evaluation: Measuring technical skill in a world where AI writes the code. The old proxies no longer hold and the new ones have not been defined yet.

Your focus will shift across these depending on where the highest-leverage work is at any given time.

What you will do

• Design and ship production ML systems across Chakra, integrity, and evaluation domains.
• Own the full ML lifecycle: problem framing, data strategy, experimentation, deployment, and iteration.
• Build evaluation infrastructure and benchmarking pipelines that reliably measure model quality before and after deployment.
• Define the architecture and production bar for different signal categories from scratch. 
• Mentor and support junior ML engineers, helping shape their technical thinking and raise the quality bar across the team.
• Establish ML best practices for the team: monitoring, model feedback loops, and quality standards.

Who you are

• 4+ years building and shipping ML systems that run in production at scale.
• Systems thinking comes naturally. Model accuracy, data pipelines, serving infrastructure, and customer outcomes are one problem, not four.
• Evaluation methodology matters as much as model performance. A metric measured wrong is worse than no metric.
• Proficient in Python, with practical experience building data pipelines and deploying models to production.

Even better if you have

• Experience with multimodal systems: vision, NLP, audio, or behavioral signal pipelines.
• LLM experience: fine-tuning, RLHF, or multi-turn agentic systems.
• Background in adversarial ML, fraud detection, or anomaly detection.
• Publications or open-source contributions in detection, robustness, or evaluation methodology.

You will thrive here if

• Messy, undefined problems are more interesting to you than optimizing within clean ones.
• Ambiguity energizes you, especially when the right framing is itself part of the work.
• Direct access to leadership, fast feedback loops, and genuinely unsolved problems is what you are looking for.
• Defining what a system should be is more compelling than maintaining what already exists.