Head of Data Science · AI & Data Science Leader · Graph AI Researcher · Generative AI · CUDA · Open Source
Building scalable AI systems across Graph AI, Generative AI, temporal machine learning, GPU-accelerated computing and production-grade MLOps.
I am an AI and Data Science leader with 20+ years of experience translating advanced research into scalable, production-ready systems. My work spans Graph AI, temporal learning, Generative AI, GPU-accelerated analytics and reliable machine-learning platforms. I combine research, hands-on engineering, published work and open-source contribution to build reproducible AI systems designed for real-world scale. I have led production-scale AI and graph analytics initiatives across large, high-volume digital ecosystems, spanning fraud intelligence, anomaly detection, federated AI and responsible AI governance.
- Led AI initiatives for one of the world's largest real-time digital payments ecosystems.
- Built production AI systems for fraud detection, mule detection, graph intelligence, federated AI and synthetic data.
- Published peer-reviewed research in Graph AI and temporal graph analytics.
- AI and ML: Graph Neural Networks, temporal graph learning, Generative AI, LLMs, Agentic AI, RAG and reinforcement learning
- Scalable computing: CUDA, NVIDIA RAPIDS, cuGraph and high-performance analytics
- Engineering: Python, SQL, C++, Docker, Kubernetes, CI/CD, MLOps, LLMOps, observability and testing
- Applied modelling: Time series, anomaly detection, incremental learning and knowledge distillation
- Open contribution to NVIDIA RAPIDS cuGraph (PR #5584) — proposes multi-seed
ego_graphoffset handling in the Python API, with regression tests and compatibility validation
| Repository | Focus and Differentiation |
|---|---|
| Topology-Aware Temporal Graph Learning | Reference implementation for topology-aware temporal node classification with reproducible benchmarks, chronological evaluation, and deterministic synthetic graph generation. |
| Cross-Modal Knowledge Distillation | ANN-to-SNN knowledge distillation framework for imbalanced tabular classification using spike encoding and hybrid distillation losses. |
| Time Series | Reproducible Python forecasting benchmark with walk-forward validation, leakage-safe backtesting, classical models, lag-based machine learning, tests, and CI. |
| Outlier Detection Tutorials | Reproducible Python tutorials and benchmarks covering statistical, distance-based, density-based, isolation, kernel, ensemble, and autoencoder methods, with exercises, tests, and CI. |
My research covers graph machine learning, temporal graphs, scalable AI, incremental learning, knowledge distillation, anomaly detection, reinforcement learning and high-performance computing. My published work includes research presented through IEEE and Springer venues.
Scalable temporal motif mining for large-scale transaction networks and high-performance graph analytics.
Adaptive fraud detection using meta-learning, Kolmogorov-Arnold Networks and ensemble strategies for evolving fraud patterns.
More research: Google Scholar · ORCID · ResearchGate · OpenReview · DBLP · Semantic Scholar · ACM Digital Library
- Book: Machine Learning for Finance: Beginner's Guide to Explore Machine Learning in Banking and Finance — BPB Publications, 2021
- Course: Data Analysis for Business and Finance — statistics, regression, time series and applied analytics
- Technical writing: Medium
- Scalable Graph AI and temporal graph learning
- GPU-accelerated graph analytics using CUDA and NVIDIA RAPIDS
- Production AI systems for fraud intelligence, anomaly detection and trustworthy AI