New opportunity

CT-NITRO-Experimental AI Engineer

EY · Kochi, India

About this role

EY is seeking an AI engineer in Kochi to provide technical leadership in developing and deploying generative AI solutions, including large language models. The role combines AI product delivery, data science, production integration and team leadership. The listed salary is competitive.

Responsibilities

Architect scalable data and machine learning pipelines and oversee robust, efficient and secure integration of AI models into production environments.
Deliver data science insights and cognitive services using complex algorithms, machine learning and AI. Develop sophisticated algorithms and improve existing machine learning techniques.
Work with business teams to identify opportunities, define requirements and success criteria, deliver solutions and turn data into recommendations that improve business decisions.
Lead cross-functional work with data scientists, engineers and product managers, managing timelines and resources. Guide AI initiatives, keep pace with advances in the field and mentor team members; the role may involve leading medium-sized projects or training junior colleagues.

Required experience and capabilities:

12+ years of experience across machine learning, data mining, predictive and statistical modeling, and large-scale acquisition, transformation and cleaning of structured and unstructured data.
Python proficiency for data wrangling, visualization and modeling; experience with ML and NLP algorithms, supervised and unsupervised learning, and deep learning methods such as recurrent and recursive neural networks, LSTMs and autoencoders.
Experience with machine learning frameworks such as scikit-learn, Keras, TensorFlow, MXNet, PyTorch or MLlib; Hugging Face Transformers and other open-source NLP/NLG modules; generative LLM fine-tuning and prompt engineering. The posting also seeks Python expertise and experience developing and deploying web apps.
A track record of leading and collaborating on strategic initiatives for at least the last 2–3 years. Strong analytical, communication, presentation, collaboration and problem-solving skills; ability to work independently and explain technical concepts to different audiences.

Preferred or additional qualifications: A bachelor’s, master’s, postgraduate or doctoral qualification in a listed field such as AI, analytics, data science, computer science, engineering or IT is preferred. Docker knowledge, Git and modern development workflow experience, and strategic consulting experience are pluses.

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

Skills for this role

Generative AILarge Language ModelsMachine LearningData ScienceData MiningPredictive ModelingStatistical ModelingData WranglingData VisualizationPythonNatural Language ProcessingNatural Language GenerationSupervised LearningUnsupervised LearningDeep LearningRecurrent Neural NetworksLong Short-Term Memory NetworksRecursive Neural NetworksAutoencodersscikit-learnKerasTensorFlowMXNetPyTorchMLlibHugging Face TransformersLLM Fine-TuningPrompt EngineeringData PipelinesDocker Web Apps Development

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

Lead generative AI delivery
The role provides technical leadership for developing and deploying generative AI solutions, including LLMs, aligned with business objectives. Prepare to explain how you would choose and assess an approach, including when fine-tuning or prompt engineering is appropriate.
Build production-ready AI pipelines
EY calls for scalable data and machine learning pipelines and robust, efficient, secure model integration into production. Prepare a concrete architecture walkthrough covering data preparation, deployment decisions and operational safeguards.
Turn business needs into measurable outcomes
The engineer works with business areas to identify opportunities, define requirements and success criteria, and turn data into recommendations. Prepare to show how you would connect a technical result to a business decision.
Deliver substantive data science work
Responsibilities include delivering cognitive services, developing algorithms and improving machine learning techniques across structured and unstructured data. Prepare to explain the reasoning and evidence behind a modeling improvement, not just the tools used.
Coordinate and mentor across disciplines
The posting assigns leadership of cross-functional teams, timelines and resources, alongside mentoring; it also seeks a track record on strategic initiatives. Prepare an example that makes technical decisions, coordination and communication to different audiences clear.

Likely challenges to think through

Balancing experimentation with production constraints
A likely challenge is deciding when an LLM approach is sufficiently useful to justify its integration costs, while meeting the posting’s scalability, efficiency and security expectations. Prepare a decision framework that compares approaches against defined success criteria and deployment constraints.
Agreeing on success across business and technical teams
Business stakeholders may value recommendations while engineers need testable requirements; those aims can be hard to reconcile. Prepare to describe how you would agree on success criteria early and communicate limitations without losing the decision-making goal.
Keeping delivery focused as AI methods evolve
The role combines project timelines and resource allocation with staying current on AI advances. A likely trade-off is whether to adopt a new method or deliver with an established one. Prepare to explain how you would evaluate potential gains, implementation effort and delivery risk.

Practice your answers

Technical practice

Suggested question: How would you design a scalable data and machine learning pipeline?

What to coverExplain how you would acquire, transform and clean large-scale structured and unstructured data before modeling. Discuss how you would check data quality and balance scalability, efficiency and secure production integration.

Suggested question: How would you turn a business opportunity into a generative AI solution?

What to coverStart with the business requirement and success criteria, then explain how you would choose an LLM-based approach and plan delivery. Address how you would test whether the resulting product serves the stated business objective.

Suggested question: When would you use prompt engineering rather than fine-tuning an LLM?

What to coverCompare the approaches in terms of available data, expected improvement, implementation effort and evaluation. Use a hypothetical business requirement rather than implying that EY uses a particular approach.

Suggested question: How would you evaluate an NLP or generative AI solution?

What to coverExplain how you would define measures tied to the agreed success criteria and assess output quality alongside practical production constraints. Suggested preparation: describe how you would examine failures and decide whether to revise data, prompts or the model.

Suggested question: How do you select and improve a predictive modeling approach?

What to coverContrast supervised and unsupervised methods for different data and business questions, including when statistical modeling is useful. Explain how you would establish a baseline, evaluate results and justify a more sophisticated algorithm.

Suggested question: Where do recurrent networks, LSTMs, autoencoders and Transformers fit?

What to coverExplain the problems these methods can address and the trade-offs that would guide selection for an NLP or other machine learning task. Keep the comparison tied to data characteristics and the required outcome, rather than assuming one architecture is always preferable.

Suggested question: How would you troubleshoot a drop in model quality?

What to coverWalk through how you would check data acquisition, transformations, modeling choices and evaluation results to isolate the cause. Suggested preparation: explain how you would prioritize a fix and verify that it improves the agreed success criteria.

Suggested question: How would you integrate an AI model into a production application?

What to coverDescribe the interface between a Python model and a deployed web app, with attention to robust, efficient and secure integration. Discuss how you would test the integration; Docker and Git can be mentioned as optional tools, not assumed requirements for a particular project.

Suggested question: How would you turn Python-based analysis into a business recommendation?

What to coverExplain how you would use data wrangling, visualization and modeling to investigate a decision. Show how you would communicate the evidence, limitations and recommended action to a business audience.

Behavioural practice

Suggested question: Tell me about a time you aligned an AI initiative with business needs.

What to coverUse situation, action and result to show how you worked with stakeholders to define requirements and success criteria. Explain how you handled any gap between the initial request and what the data could support.

Suggested question: Describe a cross-functional project you led under delivery constraints.

What to coverSet out the situation, your actions with data scientists, engineers or product managers, and the result. Focus on how you handled timelines, resources and trade-offs rather than only the technical outcome.

Suggested question: Give an example of explaining a difficult technical decision to a non-technical audience.

What to coverDescribe the situation, how you adapted your explanation and what happened afterward. Make clear how you communicated uncertainty or limitations while helping the audience make a decision.

Suggested question: Tell me about a time you mentored someone or raised a team's capability.

What to coverUse situation, action and result to explain the need you identified and the guidance you provided. Distinguish your contribution from the other person's work and describe the observable outcome.

Suggested question: Describe a tough decision you made independently in a fast-paced project.

What to coverExplain the situation, the evidence and competing priorities you considered, the action you took and the result. Include how you took responsibility and communicated the decision to affected colleagues.

Sharpen your knowledge

Topics to prepare

Generative AI and LLM design
Posting: The role provides technical leadership for generative AI solutions, including LLMs, and calls for fine-tuning and prompt engineering. Revision exercise: compare when you would use prompting versus fine-tuning for a hypothetical business task, explaining trade-offs and how you would judge success.
Statistical and machine learning methods
Posting: Experience spans predictive and statistical modeling, data mining, and supervised and unsupervised learning. Revision exercise: choose one predictive and one unsupervised approach for a hypothetical dataset, then explain their assumptions, limitations and evaluation choices.
Deep learning and NLP architectures
Posting: The role calls for ML/NLP experience and in-depth knowledge of recurrent and recursive neural networks, LSTMs and autoencoders. Revision exercise: sketch how each architecture processes data and explain where a Transformer-based approach may be more suitable.
Scalable, secure production integration
Posting: The engineer architects data and ML pipelines and oversees robust, efficient and secure integration of models into production. Revision exercise: draw a pipeline from data acquisition and cleaning through model deployment, and identify likely scale, reliability and security decisions.
Business-led AI delivery
Posting: The role works with business areas to define opportunities, requirements and success criteria, and turns data into recommendations. Revision exercise: take a hypothetical AI opportunity and state its business question, measurable success criteria, technical approach and decision-ready recommendation.
Technical leadership
Posting: The role leads cross-functional teams, manages timelines and resources, guides initiatives and may mentor junior colleagues. Revision exercise: prepare a concise account of how you would coordinate data scientists, engineers and product managers when delivery priorities conflict; treat the scenario as interview practice, not a stated EY process.

Skills to revise

Python data preparation and visualization
Posting: Python proficiency is specified for wrangling, visualization and modeling, including large-scale structured and unstructured data. Revision exercise: clean a sample of both data types in Python, document quality issues and produce a visualization that supports a modeling decision.
Modeling with a listed ML framework
Posting: Experience with frameworks such as scikit-learn and PyTorch is requested. Revision exercise: use one listed framework to train and evaluate a supervised model, then explain an error analysis and one justified improvement. The exercise is preparation, not a prescribed EY test.
Hugging Face and LLM experimentation
Posting: Experience with Hugging Face Transformers, open-source NLP/NLG modules, LLM fine-tuning and prompt engineering is requested. Revision exercise: build a small prompting comparison with a Transformers model, define evaluation criteria and outline a fine-tuning experiment without claiming either approach will be required for a specific project.
Production-facing Python web app
Posting: The role seeks Python expertise and experience developing and deploying web apps alongside production AI integration. Revision exercise: build a small Python web app that invokes a model, handles invalid input and records how you would check response reliability before deployment.
Technical and business communication
Posting: The engineer must explain difficult concepts to technical clients and non-technical audiences, present findings and collaborate across teams. Revision exercise: present the same model result twice—once with technical assumptions and limitations, and once as a brief business recommendation tied to success criteria.

Your preparation checklist

Prepare a generative AI project walkthrough
Choose a project you can discuss and outline the business objective, your technical decisions, use of prompting or fine-tuning, evaluation results, and deployment lessons. The posting emphasizes LLM delivery aligned with business goals.
Sketch a production pipeline design
Prepare a hypothetical design covering structured and unstructured data ingestion, cleaning, model integration, scalability, security, and monitoring. Be ready to explain trade-offs rather than imply this is an EY project.
Gather evidence of hands-on ML work
Select examples showing Python data wrangling and modeling, ML or NLP algorithms, and frameworks you have actually used. Prepare to explain algorithm choices and improvements; the posting also names deep-learning methods and Hugging Face Transformers.
Rehearse a business-facing discovery exercise
Practice turning an example business opportunity into requirements, success criteria, an analysis plan, and a clear recommendation. Explain the same technical decision to both technical and non-technical audiences.
Document leadership and delivery examples
Prepare concise examples of leading strategic initiatives, coordinating data scientists, engineers, and product managers, managing timelines or resources, and mentoring where applicable. Distinguish your own decisions from the team’s work.
Prepare targeted questions for EY
Ask how this role defines success for AI products, which production-integration challenges matter most, and how responsibilities are divided across engineering, data science, product, and business teams.

Availability last checked: 4 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…