Job summary

Serve as a senior architect specializing in data science machine learning PyTorch and agentic AI designing scalable hybrid solutions that power intelligent products and services. Drive end to end architecture from experimentation to production enabling responsibly built AI capabilities that create measurable value for clients and positive impact for society.

Responsibilities

Design advanced end to end architectures for data science and machine learning solutions that use PyTorch to deliver robust and scalable predictive and generative capabilities for business critical applications.

Define technical patterns for agentic AI systems that coordinate multiple intelligent components to autonomously plan act and learn while aligning with enterprise standards and responsible AI practices.

Collaborate with product owners and data practitioners to translate complex business goals into AI solution blueprints that are feasible measurable and optimized for hybrid working models.

Guide teams in structuring data pipelines feature stores and model serving layers so that PyTorch based workloads integrate efficiently with existing platforms and governance frameworks.

Review solution designs code and experiment plans to ensure that models are reproducible well documented and ready for secure deployment across multiple environments with minimal rework.

Evaluate tradeoffs among accuracy latency interpretability and cost to recommend machine learning architectures that balance innovation with reliability and long term maintainability.





Provide technical direction on the use of agentic AI frameworks to orchestrate tasks such as retrieval augmented generation tool calling and workflow automation in a protected and auditable manner.

Optimize PyTorch training and inference setups through thoughtful choice of model architectures batching strategies and hardware configurations so that systems perform reliably during day shift operations.

Partner with platform engineers to design monitoring observability and feedback loops that track data quality model drift and agent behavior enabling continuous improvement of AI services.

Document reference architectures decision records and design guidelines that help distributed teams consistently deliver high quality AI solutions while working in a hybrid arrangement with no travel expectations.

Engage with security and compliance stakeholders to embed privacy fairness and risk controls into every data science and machine learning solution so that AI outcomes support societal trust.

Mentor engineers and data professionals in advanced PyTorch usage experiment design and production readiness practices so that the broader organization grows sustainable AI capabilities.

Align solution roadmaps with company objectives by prioritizing AI use cases that improve customer experiences streamline operations and contribute to more inclusive and efficient digital services.

Certifications Required

Preferred certifications include TensorFlow Developer Certificate or equivalent cloud AI certification such as Azure AI Engineer Associate or Google Qualified Machine Learning Engineer

📌 PyTorch , Data Science , Machine Learning , Agentic AI (Mumbai)
🏢 Cognizant
📍 Mumbai