About PaytmPaytm is India’s largest digital payments platform, handling billions of transactions across users, merchants, and financial entities. Our ecosystem generates vast amounts of time-series, network and event-driven data, making representation learning, deep learning, and large-scale AI infrastructure a crucial part of our AI strategy. We are building next-generation foundation models to drive personalization, fraud detection, credit underwriting, and merchant profiling at scale.Role OverviewWe are looking for a highly technical, hands-on Vice President - Data Science who has a proven track record of successfully leading teams, developing deep learning and foundation models from scratch, and deploying AI solutions at scale in production. The ideal candidate must have strong expertise in deep mathematics, probability, statistics, optimization, and computer science fundamentals, along with experience in leading large AI teams, delivering impactful AI projects, and scaling AI models for real-world production.This role is ideal for someone who can architect, develop, optimize, and productionize deep learning models while also mentoring and leading a high-caliber team of researchers, engineers, and AI practitioners. Key Responsibilities1. AI Strategy & Leadership in Deep Learning & Foundation ModelsLead AI innovation in representation learning for Paytm’s vast ecosystem of users, merchants, transactions, and devices.Drive the development of large-scale foundation models for financial transaction temporal and graph data.Define and execute AI roadmaps to align deep learning research with business objectives.Mentor, recruit, and scale a world-class AI team, including data scientists, ML engineers, and AI researchers.Own end-to-end AI project delivery, from research and prototyping to large-scale deployment.2. Successful Deep Learning & AI Project ExecutionLed and successfully deployed large deep learning models for fraud detection, credit underwriting, and customer profiling.Developed and scaled AI-driven personalization models to enhance user experience in financial transactions.Built real-time anomaly detection systems using deep learning for transaction security.Optimized AI models for production, reducing inference latency and computational overhead while maintaining high accuracy.Drove AI adoption across multiple teams, establishing best practices for deep learning deployment and monitoring.3. Foundation Model Development for Temporal & Event-Driven DataDesigned and trained large-scale foundation models specifically for sequential, time-series, and temporal point process data.Applied transformers, temporal point processes (TPP), and self-supervised learning to improve financial forecasting and fraud detection.Built AI-driven representations of users, merchants, and devices, enabling improved credit risk assessments and fraud mitigation.Optimized multi-modal deep learning approaches that integrate text, transaction sequences, and graph data.4. Hands-on Deep Learning & Model ProductionizationArchitected, optimized, and deployed large AI models at scale, ensuring efficient inference and real-time decision-making.Developed high-performance deep learning pipelines for real-time payments, risk modeling, and financial forecasting.Built and maintained end-to-end MLOps infrastructure, ensuring smooth deployment, monitoring, and retraining of models.Implemented state-of-the-art model compression, quantization, and distillation techniques to scale AI for billions of transactions.5. Strong Computer Science & Software Engineering FoundationsExpert in data structures & algorithms, optimizing deep learning workloads for speed and efficiency.Optimized tensor computations using low-level operations for high-performance deep learning models.Built scalable AI infrastructure, leveraging distributed computing, GPU acceleration, and cloud-based AI pipelines.Implemented robust AI monitoring systems, ensuring high availability and accuracy of deployed models.6. Graph Neural Networks & Representation Learning for Paytm’s EcosystemBuilt knowledge graphs linking users, merchants, devices, and transactions to power fraud detection and risk modeling.Developed graph-based embeddings for structured and semi-structured financial data.Implemented attention-based GNNs to extract insights from transaction graphs and network-based fraud patterns.7. Reinforcement Learning for Real-Time Decision MakingDeveloped reinforcement learning (RL) agents for risk assessment, fraud prevention, and personalized financial services.Designed policy-based and value-based RL strategies for dynamic pricing and credit underwriting.Implemented multi-agent RL for adversarial fraud detection and counter-strategies.8. MLOps & Scalable AI DeploymentEstablished fully automated CI/CD pipelines for deep learning models, ensuring smooth deployment, rollback, and monitoring.Optimized model compression, quantization, and distillation for efficient inference at scale.Deployed AI models in high-throughput, low-latency production environments, ensuring fault tolerance, scalability, and real-time inference. Qualifications & Skills AI Leadership & Project ExecutionSuccessfully led AI projects from research to production, delivering high-impact deep learning solutions.Strong ability to build, mentor, and scale AI teams, fostering an innovation-driven culture.Experience in AI governance, explainability, and bias mitigation for financial AI applications.Core Computer Science & Software EngineeringExpert in data structures & algorithms, designing scalable AI architectures.Parallel & Distributed Computing: Experience with multi-threading, CUDA, OpenMP, TensorRT.Low-level System Optimization: Optimizing cache locality, memory access, and floating-point operations for AI models.Deep Learning & AI ExpertiseHands-on expertise in deep learning, designing custom architectures beyond standard libraries.Expertise in transformers, neural TPPs, reinforcement learning, graph neural networks, and self-supervised learning.Optimizing deep learning models for inference efficiency, using pruning, quantization, and distillation.Mathematics & Statistical RigorProbability & Stochastic Processes: Understanding of random processes, Markov chains, and renewal processes.Time-Series & Event-Based Modeling: Mastery of autoregressive models, Kalman filters, and state-space models.Graph Theory & Network Science: Experience with spectral graph theory, Laplacian embeddings, and diffusion-based learning.Programming & Big Data Tech StackLanguages: Python, C++, Rust (for low-latency AI applications).Deep Learning Frameworks: PyTorch, JAX, TensorFlow.Distributed Computing: Apache Spark, Dask, Flink, Ray.Cloud & MLOps: Kubernetes, Docker, Kubeflow, MLflow, AWS/GCP/Azure.High-Performance Computing: CUDA, OpenMP, TensorRT for GPU-accelerated inference. Why Join Us?Lead AI innovation at India’s largest fintech platform.Work with massive-scale real-world data and build foundation models for financial AI.Develop cutting-edge AI solutions for real-time fraud detection, credit underwriting, and user personalization.Shape the future of AI-driven financial services in one of the most data-rich ecosystems.Competitive compensation, ESOPs, and the opportunity to impact millions of users. Who Should Apply?This role is for AI leaders who have successfully built and scaled deep learning models, have led high-impact AI teams, and have a strong foundation in computer science, mathematics, and scalable AI deployment. You must be comfortable with high-performance computing, low-latency AI deployment, and scaling AI for billions of transactions.