Model Developer Python

Job Title Model Developer (Python & R) AML / Sanctions / Fintech Crime

Experience: years

Location: [Hyderabad / Mumbai / Chennai]

Employment type: Full-time

Role overview

Were looking for a hands-on Model Developer to design, build, validate and productionize statistical & ML models used for Anti-Money Laundering, sanctions screening, transaction monitoring and other fintech-crime detection use cases. The role sits at the intersection of data science, compliance and engineering you'll deliver robust, explainable models that reduce false positives, detect novel typologies and integrate into real-time and batch monitoring pipelines.

Key responsibilities

• Design and implement statistical and machine-learning models (supervised, unsupervised, graph-based, anomaly detection) for transaction monitoring, name-screening / fuzzy matching, entity resolution and typology detection.
• End-to-end model lifecycle: data exploration, feature engineering, training, hyperparameter tuning, backtesting, performance evaluation, validation and deployment.
• Build and maintain scoring pipelines for both batch and low-latency/real-time environments.
• Implement name-matching / fuzzy matching algorithms (e.g., Levenshtein, Jaro-Winkler, phonetic methods), sanctions list handling and multi-lingual normalization.
• Work with compliance/ops teams to translate business rules and typologies into model features and detection logic.
• Create model explainability and documentation (model cards, assumptions, limitations, versioning) to support model governance and audit.
• Monitor model drift, performance metrics, false positive/false negative rates and run periodic recalibration/retraining.
• Participate in model validation and support internal/external audits and regulatory inquiries.
• Collaborate with data engineers and SRE/DevOps to containerize models (Docker), CI/CD for models, and ensure secure deployment in cloud/on-prem.
• Communicate technical results and tradeoffs clearly to non-technical stakeholders.

Required technical skills & experience

• 4–12 years hands-on experience building and deploying models in production, specifically in AML, sanctions screening, fraud or fintech crime domains.
• Strong programming in Python and R (both languages used for modeling and analysis). Proficiency with libraries such as scikit-learn, XGBoost/LightGBM/CatBoost, pandas, statsmodels in Python and in R.
• Experience with unsupervised methods (clustering, isolation forest, autoencoders), graph analytics (network features / GNN familiarity a plus) and time-series / sequence modeling.
• Expertise in name matching and fuzzy matching algorithms, entity resolution, and rule augmentation for sanctions screening.
• Strong SQL skills and experience with large transactional datasets.
• Exposure to big-data technologies (Spark / PySpark) and columnar stores is desirable.
• Experience deploying models: Docker, REST APIs, batch schedulers (Airflow, Luigi), and familiarity with cloud platforms (AWS / GCP / Azure).
• Understanding of model governance, validation, explainability (SHAP/LIME), and regulatory requirements for AML/KYC.
• Version control (Git) and unit testing for analytics code.

Preferred / Nice-to-have

• Prior experience in a regulated bank, payments company, or fintech fraud/AML team.
• Working knowledge of sanctions lists (OFAC, UN, EU, HMT) and PEP lists, and tradecraft for handling list updates & fuzzy matching across scripts/languages.
• Familiarity with graph databases (Neo4j) or libraries for graph analytics (NetworkX, StellarGraph).
• Certifications: CAMS, CISSP, FRM or relevant compliance/anti-money-laundering credentials.
• Experience with production monitoring tools and MLOps frameworks (MLflow, Seldon, BentoML).

Behavioral & communication skills

• Strong problem solving and investigative mindset — comfortable tracing alerts back to root causes / data issues.
• Able to explain model behaviour and limitations to compliance, legal and business stakeholders.
• Team player: works cross-functionally with product, data engineering, compliance and operations.
• Pragmatic: balances statistical rigor with operational constraints (false positive reduction vs. detection sensitivity).

Education

• Bachelor's or Master's in Computer Science, Statistics, Mathematics, Data Science, Economics, or related quantitative field (or equivalent practical experience).

Kindly note: We are looking for candidates to join with a notice not max than 30 days.