New opportunity

D&T Data Scientist II

General Mills · Mumbai, India

About this role

General Mills’ Digital and Technology India team seeks a Data Scientist II to build enterprise intelligent automations and machine-learning solutions. Use cases include demand forecasting, text classification, operational analytics, logistics and price optimization, ecommerce, and market mix modelling. The role is based in Powai, Mumbai, MH, with a hybrid working arrangement.

Responsibilities

Partner with Lead Data Scientists and business stakeholders to define problems, deliver useful insights and intelligent solutions, and drive measurable outcomes.
Develop maintainable, scalable AI solutions. Work with ML and Systems Engineers to deploy models efficiently, monitor production performance, and address model degradation.
Evaluate alternative approaches, apply sound judgment to varied problems, manage assigned priorities, and build domain knowledge and stakeholder relationships.
Help develop data science standards, documentation, team learning, and a community of practice. Track developments in data science, machine learning, and Agentic AI and assess their business relevance.

Minimum qualifications:

Bachelor’s degree in Technology, Engineering, or Science and 6–8 years of total analytics experience.
Expertise in supervised machine-learning algorithms, regression, decision trees, ensemble models, time series, forecasting, and neural networks; exposure to unsupervised learning and NLP.
Proficiency in Google Cloud Platform, SQL, Python, and R. Experience constructing and solving linear, mixed-integer, constraint, and nonlinear programming problems; ability to design or implement heuristic algorithms. Experience with problems such as network optimization, cutting stock, and knapsack.
Experience with Agile development using sprints, point estimation, and daily stand-ups. Prior work with at least a couple of major commercial or open-source operations-research tools, such as Google OR-Tools, IBM ILOG CPLEX, Gurobi, AIMMS, or FICO Xpress.
Domain experience in consumer packaged goods, manufacturing, retail, ecommerce, energy, sales, or commercial functions, plus strong collaboration and storytelling skills.

Preferred

Knowledge and use of agentic or generative tools. Applicants must meet the minimum age requirements in the country where the job is located.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Data ScienceMachine LearningSupervised LearningRegressionDecision TreesEnsemble ModelsTime Series ForecastingNeural NetworksUnsupervised LearningNatural Language ProcessingGoogle Cloud PlatformSQLPythonRLinear ProgrammingMixed-Integer ProgrammingConstraint ProgrammingNonlinear ProgrammingHeuristic AlgorithmsNetwork OptimizationCutting StockKnapsack ProblemsGoogle OR-ToolsIBM ILOG CPLEXGurobiAIMMSFICO XpressAgile DevelopmentCollaborationStorytelling

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Frame business problems around measurable outcomes
The role partners with Lead Data Scientists and stakeholders to clarify needs, build domain knowledge, and deliver practical AI and ML solutions with measurable value.
Choose sound ML approaches across varied use cases
The posting names demand forecasting, text classification, operational analytics, ecommerce, and market mix modelling. It calls for comparing approaches and applying judgment across problems of diverse scope.
Formulate and solve optimization problems
The qualifications emphasize linear, mixed-integer, constraint, and nonlinear programming; heuristics; and problems such as network optimization, cutting stock, and knapsack.
Deliver models that remain useful in production
The role develops maintainable, scalable solutions and works with ML and Systems Engineers on efficient deployment, performance monitoring, and model degradation.
Strengthen shared practices while executing assigned work
The role calls for disciplined priority management, Agile delivery, and contributions to data science standards, documentation, team learning, and a community of practice.

Likely challenges to think through

Balancing solution quality against operational constraints
Preparation hypothesis: across forecasting and optimization use cases, a technically strong method may be too slow, complex, or restrictive to use. Prepare to explain how you would compare approaches against business objectives, feasibility, and implementation cost rather than model metrics alone.
Keeping deployed models effective as conditions change
Preparation hypothesis: scalable deployment and post-deployment degradation may require trade-offs between model complexity and ease of monitoring or maintenance. Prepare an example of how you would define performance checks, investigate deterioration, and coordinate a response with engineers.
Turning varied stakeholder needs into focused delivery
Preparation hypothesis: diverse use cases and assigned priorities may compete for time while stakeholders define value differently. Prepare to describe how you would clarify a problem statement, agree on measurable outcomes, communicate trade-offs, and keep work moving within sprints.

Practice your answers

Technical practice

How would you design and evaluate a demand-forecasting solution?

What to coverDemand forecasting is a listed use case. Suggested preparation: explain how you would choose a forecast horizon, prevent time-based leakage, compare baselines with more complex models, and evaluate errors in terms of the business decision.

When would you choose regression, a tree ensemble, or a neural network for a supervised-learning problem?

What to coverThe posting calls for expertise in these approaches. Suggested preparation: compare predictive performance, data needs, interpretability, maintenance, and how you would test whether added complexity is worthwhile.

How would you approach a text-classification problem?

What to coverText classification is a listed use case, and the qualifications mention NLP and unsupervised learning exposure. Suggested preparation: discuss labels and data quality, an appropriate baseline, evaluation of class-specific errors, and whether unsupervised exploration would add value.

How would you formulate a business problem as a linear, mixed-integer, constraint, or nonlinear program?

What to coverThe role calls for constructing and solving each of these problem types. Suggested preparation: define decision variables, objective, and constraints, then explain which features of a problem determine the formulation and how you would check that its solution is practical.

When would you use an exact solver rather than a heuristic for network optimization, cutting stock, or knapsack?

What to coverThe posting names these problem areas, heuristic design, and experience with multiple operations-research tools. Suggested preparation: compare solution quality, runtime, scalability, and explain how you would validate a heuristic against a solver result when feasible.

How would you organize a reproducible data-science workflow using SQL, Python or R, and Google Cloud Platform?

What to coverProficiency in all four is listed. Suggested preparation: describe data checks, repeatable transformations, model evaluation, and handoff artifacts without assuming a particular cloud service or prescribed toolchain.

How would you deploy and monitor a model with ML and Systems Engineers?

What to coverThe role explicitly includes scalable deployment, production monitoring, and addressing model degradation. Suggested preparation: cover deployment handoffs, performance and data-quality signals, investigation of deterioration, and criteria for corrective action.

How would you evaluate a price-optimization or market-mix modelling approach?

What to coverBoth are listed use cases. Suggested preparation: clarify the business objective and available evidence, compare plausible modelling approaches, and distinguish a good model fit from a useful decision; treat any proposed measurement design as your approach, not a stated company process.

How would you assess whether an agentic or generative tool is useful for a business problem?

What to coverThe posting prefers knowledge and use of these tools and asks the role to assess relevant trends. Suggested preparation: identify a concrete use case, compare it with a simpler alternative, and discuss evaluation, reliability, and maintainability before recommending adoption.

Behavioural practice

Tell me about a time you clarified an ambiguous stakeholder problem and defined a measurable outcome.

What to coverStakeholder engagement and measurable outcomes are stated responsibilities. Suggested preparation: use situation, action, and result to show the questions you asked, how you agreed on the problem, and what evidence showed whether the work helped.

Describe a time you coordinated with engineering partners to put an analytics solution into production.

What to coverThe role involves collaboration with ML and Systems Engineers. Suggested preparation: give a situation, action, and result that distinguish your contribution from your partners’ work and explain how you resolved a deployment or maintenance trade-off.

Tell me about a time competing priorities affected your work in an Agile sprint.

What to coverThe posting mentions managing assigned priorities and Agile work with sprints, point estimation, and daily stand-ups. Suggested preparation: explain the situation, how you communicated and adjusted the plan, and the result for delivery.

Describe a time you explained a complex analysis to a business or senior stakeholder.

What to coverCollaboration, storytelling, and relationships with senior stakeholders are emphasized. Suggested preparation: use situation, action, and result to show how you adapted the explanation, communicated uncertainty, and supported a decision.

Tell me about a time you improved a team’s data-science practices or shared a new method.

What to coverThe role supports standards, documentation, team learning, and a community of practice. Suggested preparation: describe the situation, your specific action, how others used the contribution, and an observable result rather than claiming a company-wide change.

Sharpen your knowledge

Topics to prepare

Supervised learning and forecasting
The posting calls for expertise in regression, decision trees, ensembles, neural networks, time series and forecasting. For revision, compare two approaches to a hypothetical demand-forecasting problem; explain validation, error measures and when each approach would be preferable.
Optimization formulations
The role explicitly covers linear, mixed-integer, constraint and nonlinear programming, with examples including network optimization, cutting stock and knapsack problems. For revision, formulate one such problem with decision variables, objective and constraints, then explain when a heuristic might be useful.
Production model performance
The role involves working with ML and Systems Engineers on scalable deployment, monitoring and model degradation. For revision, outline how you would distinguish a data-quality issue from deteriorating model performance and decide what to investigate next.
Unsupervised learning and NLP
The minimum qualifications specify exposure to both, and text classification is a listed use case. For revision, sketch a text-classification approach and explain where an unsupervised method might help; this is a preparation exercise, not a prescribed project.
Business framing and method selection
The posting emphasizes clarifying stakeholder needs, comparing approaches and driving measurable outcomes across domains such as consumer packaged goods, retail and ecommerce. For revision, turn a hypothetical price-optimization request into a problem statement, success measure and comparison of possible methods.
Agentic and generative tools
Knowledge and use of these tools is preferred, while the role also asks for judgment about Agentic AI’s business relevance. As optional preparation, choose a hypothetical business task and explain where such a tool might help, where it might not, and how you would evaluate its usefulness.

Skills to revise

SQL, Python and R analysis
The posting requires proficiency in all three. As a practice exercise, use a small sample dataset to write a SQL aggregation and reproduce a simple regression or forecast analysis in Python and R, documenting assumptions and results.
Operations-research tools
The posting asks for prior work with at least a couple of major OR tools and names Google OR-Tools, IBM ILOG CPLEX, Gurobi, AIMMS and FICO Xpress as examples. For revision, formulate and solve a small knapsack or network problem with two tools you know, then compare their solutions and modelling choices.
Heuristic design
Designing or implementing a heuristic is an explicit qualification. Practice writing a heuristic for a small cutting-stock or knapsack instance, then compare its result and runtime with an optimization solution.
Google Cloud Platform workflow
Google Cloud Platform proficiency is listed, without specifying services. As a suggested exercise, diagram how data, model execution and performance checks could fit into a maintainable GCP-based workflow; state the choices you would need to confirm rather than assuming a required architecture.
Stakeholder storytelling
Collaboration and storytelling are explicit requirements. Practice a two-minute explanation of a forecasting or optimization result that states the business question, measurable outcome, limitations and recommended next step.
Agile delivery planning
The posting specifies sprints, point estimation and daily stand-ups. As practice, break a hypothetical analytics task into sprint-sized work, estimate it, and prepare a brief stand-up update covering progress, next steps and blockers.

Your preparation checklist

Build a requirements evidence sheet
List concrete examples for the posting’s analytics experience, degree, domain work, ML methods, optimization work, tools, and Agile practices. Mark what each example demonstrates without claiming experience you cannot substantiate.
Rehearse an ML project walkthrough
Choose a real forecasting, classification, or comparable project. Prepare a concise account of the business problem, data, baseline, model alternatives, validation, results, and decisions made from the findings.
Practice an optimization design exercise
Using a hypothetical logistics or price-optimization problem, define the objective, decision variables, and constraints. Explain when you would use a mathematical program or heuristic, and be ready to discuss two OR tools you have actually used.
Prepare technical work samples
Select examples that show your use of SQL, Python, R, and Google Cloud Platform, as applicable. Be ready to explain data handling, implementation choices, and how you checked correctness; distinguish hands-on work from familiarity.
Outline a production model plan
Sketch how you would work with ML and Systems Engineers to deploy a maintainable model, monitor performance, detect degradation, and decide when to investigate or update it. Identify what you would document.
Prepare a stakeholder and delivery story
Use a real example to show how you clarified a business need, compared approaches, communicated an outcome, and managed priorities. Include your contribution to sprint planning, estimation, stand-ups, or team learning where relevant.
Draft focused questions for the team
Ask which use cases are priorities, how outcomes are measured, which OR tools are used, and how deployment and monitoring responsibilities are shared. You could also ask where agentic or generative tools have a practical role; the posting lists their use as preferred.

Availability last checked: 5 Nov 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…