ABOUT THE ROLE:

As Staff Data Scientist for the Fraud Risk area, you will work closely with the Business Team, Product Managers, Data Governance team, Analysts, Scientists, and Data Engineers in order to deliver Company, Business, and Product OKRs. You will be at the forefront of addressing fraud globally for Tide, utilizing your deep expertise in classical Machine Learning to build and continuously improve models that detect and mitigate fraud.

This role is an Individual Contributor (IC) position. You will spend most of your time dealing with highly imbalanced data use cases, data drift in pipelines, and identifying rapidly changing fraud patterns using classical ML. Some of your time will also focus on implementing and scaling production-ready GenAI and Agentic AI workflows. You will act as the Subject Matter Expert (SME) across the team, improving our technical standards.

As a Staff Data Scientist you ll be:

- Design and develop advanced predictive ML models tailored for global fraud detection, risk assessment, and anomaly detection.
- Design and develop production-grade GenAI and Agentic AI solutions (leveraging frameworks like LangGraph on AWS or GCP) to automate alert handling and accelerate fraud investigations.
- Tackling highly imbalanced datasets using advanced techniques, ensuring models optimize for real business impact (minimizing false positives while catching evolving fraud).
- Proactively identifying, measuring, and resolving data drift and concept drift in production pipelines to maintain continuous model performance.
- Presenting compelling insights and demonstrating strong storytelling skills to translate complex model outputs for non-technical stakeholders.
- Acting as the definitive technical IC and SME for the team setting technical standards and reviewing architectures to push the boundaries of what 'best' looks like.




- Collaborate with ML engineers to deploy models, establish CI/CD pipelines, and implement robust model tracking and observability.
- Partnering with Data Engineering to optimize feature engineering and feature stores for real-time and batch fraud decisioning.
- Working with Product and Business teams to translate complex fraud typologies and business requirements into rigorous mathematical formulations and data science problems.
- Build and track metrics for the performance of our models and their direct impact on the business bottom line, feeding this back to Product and Business Teams.
- Ability to deal with ambiguity and propose innovative solutions without getting blocked.

WHAT WE ARE LOOKING FOR:

- You have 10+ years of experience in Data Science or Machine Learning, with a substantial portion of your career dedicated to fighting fraud or risk mitigation in a fast-paced environment.
- Individual Contributor: You are highly comfortable acting as a dedicated, fully hands-on technical IC and leading by example.
- Absolute SME in Classical ML: Deep, theoretical, and practical understanding of classical machine learning algorithms (e.g., XGBoost, LightGBM, Random Forests, SVMs, Ensemble methods) and statistical modeling.
- Deep Learning Synthetic Data: Strong experience building adversarial models for fraud detection and working with synthetic data generation to bolster model robustness.
- Production GenAI Agentic AI Experience:



Hands-on experience taking GenAI and multi-agent systems to production (using tools like LangGraph, AWS Bedrock, or GCP ecosystem) that have delivered real, measurable ROI.
- Technical Eminence: Demonstrated technical inclination, such as submitting or publishing research papers at technical conferences, holding patents, or making significant open-source contributions, is highly advantageous.
- Mastery of Imbalanced Data: Extensive hands-on experience handling highly imbalanced datasets, utilizing appropriate sampling methods, cost-sensitive learning, and relevant evaluation metrics (Precision-Recall AUC, F1-score, custom loss functions).
- Drift Observability: Strong hands-on experience detecting and addressing data drift, concept drift, and model degradation in production systems.
- Solid Engineering Background: You have good technical knowledge in SQL, strong in Python programming (including exposure to PyTorch and Hugging Face), and a good understanding of performance optimization in the end to end data pipeline including ML/DS inferencing.
- You have a high level understanding of big-data technologies such as Spark, Hadoop etc. Strong knowledge of Cloud (AWS or GCP).
- You re a self-starter who can work comfortably in a fast-moving company where priorities can change and processes may need to be created from scratch with minimal guidance.

OUR TECH STACK

(You don't have to excel in all, but willing to learn them):

- Databricks on AWS/GCP
- Python
- Snowflake/Big Query
- Tecton/Databricks - feature store
- Fiddler - model observability platform

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.

📌 Staff Data Scientist (Fraud & Risk) (Hyderabad)
🏢 Tide Software
📍 Hyderabad