Note: The job is a remote job and is open to candidates in USA. HackerRank is a company that helps enterprises like NVIDIA, Amazon, and Microsoft hire and upskill developers based on skills. They are seeking a Machine Learning Engineer focused on integrity to standardize model quality across integrity signals and drive strategy-level decisions in the development of new signals.

Responsibilities

• Standardize how model quality is defined, measured, and reported across all integrity signals
• Build the evaluation infrastructure, golden datasets, and benchmarking pipelines that give us and our customers genuine confidence in what we ship
• Own the performance improvement strategy for each signal
• Explore newer architectures, emerging research, and different training paradigms
• Define the ML strategy for new signals from scratch: audio analysis, gaze tracking, behavioral anomalies
• Continuously monitor how assessment fraud tooling is evolving
• Evaluate new models as they emerge
• Know when to abandon a strategy that is no longer moving the needle
• Systematically surface edge cases, build training data around them, and turn every customer-reported failure into a model that is harder to fool
• Drive strategy-level decisions: which new signals to build, whether to use models at all, and what the evidence says

Skills

• You have shipped ML systems in production that real users and real businesses depend on
• You have deep intuition for where precision leaks happen and how to find them systematically, not by luck
• You think in systems. A signal's accuracy number, its data pipeline, its serving infrastructure, and its customer-facing outcome are one problem to you
• You care as much about evaluation methodology as model performance. You know that a metric measured wrong is worse than no metric
• You are genuinely curious about adversarial dynamics. The fact that your model will be attacked is interesting to you, not exhausting
• Standardize how model quality is defined, measured, and reported across all integrity signals
• Build the evaluation infrastructure, golden datasets, and benchmarking pipelines that give us and our customers genuine confidence in what we ship
• Own the performance improvement strategy for each signal. Explore newer architectures, emerging research, and different training paradigms
• Define the ML strategy for new signals from scratch: audio analysis, gaze tracking, behavioral anomalies
• Continuously monitor how assessment fraud tooling is evolving. Evaluate new models as they emerge
• Systematically surface edge cases, build training data around them, and turn every customer-reported failure into a model that is harder to fool
• Drive strategy-level decisions: which new signals to build, whether to use models at all, and what the evidence says
• Experience with multimodal systems in production: vision, audio, or behavioral signal pipelines
• Background in adversarial ML or fraud/anomaly detection
• Publications or open-source work in detection, robustness, or model evaluation
• Prior experience defining what production-ready means for a new signal category from scratch

Company Overview

Company H1B Sponsorship