AI Architect

Responsibilities:

• Architect and build AI-powered platforms and intelligent systems for ONDC.
• Design, train, fine-tune, distil, and deploy Small Language Models (SLMs) for domain-specific commerce intelligence and reasoning.
• Lead model training and fine-tuning pipelines, including dataset preparation, experimentation, hyperparameter optimisation, evaluation, and performance tuning.
• Build and optimise LLM/SLM-based systems, focusing on model architecture, transformer-based networks, inference optimisation, and production readiness.
• Develop deep knowledge and reasoning systems combining structured, unstructured, and real-time commerce data.
• Design intelligent agent orchestration frameworks for buyers, sellers, and ecosystem participants.
• Drive production deployment of trained and fine-tuned models, ensuring scalability, low latency, reliability, and cost efficiency.
• Build scalable AI infrastructure and model-serving pipelines suitable for India-scale adoption.
• Contribute hands-on through architecture, coding, model experimentation, code reviews, and end-to-end production deployment.

Requirements:

• 8-12 years of strong software engineering and research experience.
• Proven track record of building large-scale distributed systems and advanced AI platforms.
• Hands-on experience with Small Language Models (SLMs), including model training, fine-tuning, distillation, and deployment.
• Strong experience in LLM development and customisation, including training or fine-tuning models for specific domains and use cases.
• Strong understanding of neural networks, transformers, model training methodologies, knowledge systems, and inference optimisation.
• Experience building and deploying end-to-end machine learning or language model pipelines, from data preparation and training to evaluation and production serving.
• Experience designing intelligent orchestration systems, AI agents, or complex decision-making platforms.
• Ability to reason deeply about scale, latency, reliability, accuracy, privacy, and infrastructure cost.
• Strong ownership, curiosity, research orientation, and a hands-on builder mindset.