AI/ML Professional Build the intelligence layer that turns open network data into action for India's digital commerce. New Delhi | 7-13 years | Full-time | Data Science & Analytics About ONDC ONDC is building an open network for digital commerce in India — enabling buyers, sellers, logistics providers and technology platforms to connect and transact across an interoperable network. We are working at the intersection of technology, commerce, regulation and public digital infrastructure. The problems are complex, the ecosystem is evolving, and many of the challenges do not come with an established playbook. That is where you come in. The Role ONDC is building an intelligence layer: insights that drive actions at the central gateway layer to improve technical, operational and business efficiency across the network. We are looking for an AI/ML professional to build recommendation models across domains that produce such actionable insights and drive measurable business impact. In this role, you will use traditional algorithms for initiatives such as demand forecasting, channel margin optimization, customer segmentation and churn analysis. You will also need experience deploying production-grade ML systems spanning NLP, deep learning, predictive modeling, agentic AI, RAG pipelines and LLM fine-tuning. You will collaborate closely with product, engineering and business teams, and your work will directly affect millions of buyers and hundreds of thousands of small sellers across the country. This is a hands-on, individual-contributor position suited to someone with a few years of applied experience who is ready to own problems end-to-end — from framing and data wrangling through modeling, evaluation and working with engineers to ship. What You'll Own Build and evaluate models across one or more network problem areas: search relevance and ranking, product and price recommendations, demand forecasting, logistics ETA and matching, fraud and quality, or pricing intelligence. Work with partial, distributed data: design analyses and models that are robust to the fragmented visibility inherent in a decentralized network, where no single participant sees the full buyer-to-delivery journey. Anomaly detection framework: use an agentic framework to identify anomalies, trace the drivers behind them, and create workflows with recommended actions for the operations and business teams; incorporate a feedback loop for recursive learning. Ship responsibly: partner with ML and platform engineers to productionize models, and uphold the network's commitments to privacy, fairness and explainability, especially where models affect small sellers. Monitor and improve: track model and network health, investigate failure points in the transaction funnel, and iterate based on observed outcomes. What Success Looks Like You become the person who turns ONDC's network data into recommendations the business acts on — models that reach production, hold up under fragmented visibility, and deliver measurable impact for buyers, sellers and network operations. What We're Looking For Master's degree in Computer Science, Statistics, Mathematics, Economics or a related quantitative field. 8+ years of applied data science experience delivering models or analyses that shipped to production or directly informed decisions. Strong programming skills in Python and fluency with the data/ML stack (e.g., pandas, NumPy, scikit-learn, and at least one deep-learning framework such as PyTorch or TensorFlow). Solid SQL skills and comfort working with large datasets and distributed data tools. Sound grounding in statistics, machine learning fundamentals and rigorous model evaluation. Experience with experimentation (A/B testing) and an understanding of when correlation is not causation. Experience with projects using techniques such as logistic regression, decision trees, gradient boosting, k-means clustering, recommendation models and time series forecasting. Preferred: experience in e-commerce, marketplaces, logistics, fintech or other large-scale consumer platforms. Preferred: hands-on work in one or more of search/ranking, recommender systems, time-series forecasting, fraud/anomaly detection, or operations research/optimization. Preferred: familiarity with NLP and entity resolution, particularly for multilingual or noisy text. Preferred: exposure to building agentic frameworks is a plus, but only if it was a meaningful project that you can explain. This Role Is For You If... You enjoy turning fragmented, distributed data into recommendations that teams act on. You want to own problems end-to-end, from framing and data wrangling through modeling, evaluation and shipping with engineers. You care about building models that are private, fair and explainable, especially where they affect small sellers. Why ONDC? This is an opportunity to work on open digital infrastructure at national scale — at the intersection of technology, commerce and policy. You won't just do a job. You will help shape the ecosystem it is building. Read more about ONDC (click to read) Visit our website at ondc.org McKinsey Report Our latest newsletter
