Data Specialist

Title - Data Science Specialist (AI / ML)

Location - Pune

Experience : 4 - 10 years of experience

Hybrid Work Model (3 days WFO in a week)

Notice Period - Immediate to 30 days max

Job Title:

Senior Data Scientist – Fraud and Anomaly Detection

About the Role:

We are seeking an experienced Senior Data Scientist specializing in Fraud and Anomaly Detection to join our dynamic team. In this pivotal role, you will develop and deploy advanced models and algorithms to identify fraudulent activities and anomalies within complex datasets. Your expertise in Graph Anomaly Detection and Graph Neural Networks will be crucial in enhancing our detection capabilities.

Key Responsibilities:

- Design, develop, and implement sophisticated machine learning models tailored for fraud and anomaly detection using cutting-edge techniques.
- Implement Graph Neural Networks (GNNs) with PyTorch Geometric to capture complex relationships and dependencies in graph-structured data.
- Develop and refine algorithms for Graph Anomaly Detection, focusing on node, edge, and subgraph anomalies.
- Apply various GNN architectures such as GCN, Graph

SAGE, GAT, GINE, and Hetero Graph Learning.
- Leverage NLP techniques and models like BERT, Text Classification, Text Extraction, LLMs, and LLM Fine-Tuning for data preprocessing and feature extraction.
- Perform in-depth exploratory data analysis (EDA) and engineer features, including graph-based features like node embeddings, edge attributes, and graph metrics.
- Utilize spectral clustering and community detection methods to enhance feature representation.
- Conduct research on emerging methodologies in fraud detection, anomaly detection, and graph-based machine learning.
- Prototype, evaluate, and experiment with novel architectures and algorithms, including attention mechanisms in GNNs.
- Collaborate closely with data science and engineering teams to ensure seamless model deployment, scalability, and real-time processing.

Technical Skills &

Qualifications:

- Extensive proficiency in Python and related libraries such as PyTorch, PyTorch Geometric, Tensor Flow, Keras, Transformers, and Scikit-learn.
- Strong experience working with Graph-based Machine Learning and NLP techniques.
- Hands-on experience with graph processing frameworks like Network

X and DGL (Deep Graph Library).
- Solid understanding of machine learning paradigms: supervised, unsupervised, and semi-supervised learning.
- Deep expertise in deep learning architectures for graph data, including GAT, Graph

SAGE, and attention mechanisms.
- Experience developing models for large-scale datasets and deploying them into production environments.

Ideal Candidate Profile:

- Proven experience in fraud, outlier, or anomaly detection projects.
- Strong analytical and problem-solving skills with an innovative mindset.
- Excellent collaboration and communication skills, capable of working cross-functionally with teams of data scientists and software developers.

If you are passionate about leveraging advanced graph and NLP techniques to combat fraud and enhance anomaly detection, we would love to hear from you!