In short
AI/ML Engineer in Delhi working on fraud detection: cleaning and labeling SDK identity, device, and behavior data; building leakage-free training pipelines; developing models for account takeover, bots, and scams; and combining model outputs into trust scores with low-latency inference alongside backend and SDK engineers.
About the role
Responsibilities:
- Clean, label, and analyze SDK identity, device, and behavior data; build reproducible feature and training pipelines with time-based splits that prevent data leakage.
- Check label quality, missing signals, and consistency between training features and live inference inputs.
- Develop and compare baseline models for account takeover, bots, and scam indicators; combine model outputs and partner-engine signals into useful trust scores.
- Evaluate precision, recall, false positives, and performance across customer segments; document failure cases and develop clear reason codes for scores and decisions.
- Use confirmed fraud and false-positive feedback to improve datasets, monitor drift and retrain models, track experiments, and validate changes in shadow deployments.
- Collaborate with backend and SDK engineers on reliable, low-latency inference and partner-model integration; support liveness or deepfake evaluation where relevant as the roadmap develops.
Requirements
- Clean, label, and analyze SDK identity, device, and behavior data
- Build reproducible feature and training pipelines with time-based splits that prevent data leakage
- Check label quality, missing signals, and train/live inference input consistency
- Develop and compare baseline models for account takeover, bots, and scam indicators
- Combine model outputs and partner-engine signals into trust scores
- Evaluate precision, recall, false positives across customer segments; document failure cases and reason codes