New opportunity

AI Architect (EA100FT RM 3966)

Source-right · Bangalore, India

About this role

The AI Architect designs and implements AI and generative AI solutions that address business problems and align with organizational goals. The role spans machine learning, language models, NLP, computer vision and predictive analytics, with an emphasis on scalable, secure and maintainable architecture.

Responsibilities

Define AI architecture, frameworks and strategy; evaluate tools and platforms; and oversee or contribute to the development, training and optimization of large and small language models. Integrate solutions with business processes and comply with Arrow’s internal policies and guidelines.
Architect on-premises and cloud solutions, including Azure PaaS and SaaS deployments. Work with external vendors and partners where appropriate.
Define scope, deliverables, plans, timelines and milestones for AI initiatives and the in-house GenAI platform. Work with business units to determine ROI and project KPIs; manage delivery, risks and contingency plans.
Ensure projects meet data privacy requirements, industry standards and ethical guidelines. Coordinate research teams, data scientists, engineers, product managers, business stakeholders, and onsite and offshore AI teams.

Qualifications and requirements:

The listing specifies 10+ years of experience, a minimum of 5 years in AI projects, data science or a related field, and 4 years of work in GenAI/AI. It calls for experience with 3–4 GenAI solutions involving RAG, LLMs, fine-tuning or text-to-SQL.
Bachelor’s or master’s degree in Computer Science, Artificial Intelligence, Data Science or a related field; an advanced degree is preferred.
Expertise in AI and GenAI methods, including supervised and unsupervised learning, deep learning, NLP and computer vision. Proficiency in Python and .NET; knowledge of Azure, GCP, APIs, microservices, data engineering and cloud architecture. Relevant tools include LangGraph, LangChain, Azure AI Foundry, Ollama, Databricks, Hadoop and Apache Spark.
Technical project management, agile tools such as Jira and Trello, communication, problem-solving, and knowledge of AI ethics and regulatory compliance are required. The stated notice period is 0–30 days. The job location is listed as BangalorePune.

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 ModelsRetrieval-Augmented GenerationLangGraphLangChainFine-tuningText-to-SQLMachine LearningDeep LearningNatural Language ProcessingComputer VisionPredictive AnalyticsPython.NETMicrosoft AzureGoogle Cloud PlatformAzure AI FoundryDatabricksOllamaHadoopApache SparkAPIsMicroservicesData EngineeringCloud ArchitectureJiraTrelloAgile MethodologiesAI EthicsRegulatory Compliance

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

Design deployable AI architecture
The posting calls for scalable, secure, maintainable AI solutions across on-premises and cloud environments, including Azure PaaS and SaaS. Prepare to explain architecture and platform choices against those constraints.
Lead GenAI solution design
The role includes evaluating tools and contributing to the development, training and optimization of large and small language models. The required skills name RAG, fine-tuning and text-to-SQL solutions. Prepare concrete design decisions and evaluation criteria for these approaches.
Tie delivery to business outcomes
The architect defines scope, plans and milestones for AI initiatives and the in-house GenAI platform, and works with business units on ROI and KPIs. Prepare to show how technical deliverables can be measured against a business objective.
Build in governance and integration
Solutions must fit existing business processes and meet data privacy, industry-standard and ethical requirements; the posting also references Arrow’s internal policies. Prepare to explain how these constraints affect design and deployment decisions.
Coordinate delivery across teams
The role liaises with research, data science, engineering, product and business stakeholders, and coordinates onsite and offshore AI teams. Prepare examples of how architecture decisions and delivery risks can be communicated across those groups.

Likely challenges to think through

Choosing an appropriate deployment boundary
Likely challenge, not a confirmed employer problem: on-premises and cloud options may pull security, scalability and maintainability in different directions. Prepare a decision framework that compares data location, integration needs, operational ownership and Azure deployment options.
Proving value despite model and data uncertainty
Likely challenge, not a confirmed employer problem: data limitations or model performance could undermine a planned ROI target. Prepare to connect baseline measures and project KPIs to model evaluation, risk mitigations and contingency decisions.
Keeping distributed work aligned
Likely challenge, not a confirmed employer problem: coordinating business units, technical disciplines and onsite/offshore teams may create competing priorities or handoff delays. Prepare an approach to milestones, decision ownership and escalation that protects delivery without obscuring risks.

Practice your answers

Technical practice

How would you translate a business problem into an AI architecture?

What to coverThe role calls for AI strategies aligned with organizational goals and project-level ROI and KPIs. Suggested preparation: explain how you would define the use case, compare AI and non-AI approaches, and select measurable outcomes before choosing an architecture.

How would you design and evaluate a retrieval-augmented generation solution?

What to coverThe posting names RAG, LangChain and LangGraph among the relevant skills. Suggested preparation: cover retrieval quality, grounding, access to source data, and how you would diagnose inaccurate or unsupported answers.

When would you use a large model, a small model or fine-tuning?

What to coverThe role includes development, training and optimization of large and small language models, and lists fine-tuning experience. Suggested preparation: compare task quality, data needs, latency, cost and maintainability, including how you would evaluate whether optimization helped.

How would you make a text-to-SQL solution reliable?

What to coverText-to-SQL is one of the GenAI solution types named in the posting. Suggested preparation: discuss schema context, query validation, permissions and tests for correctness; distinguish these proposed safeguards from any employer-specific implementation.

How would you choose between on-premises and cloud deployment for an AI solution?

What to coverThe role covers on-premises architecture, Azure and GCP, with particular emphasis on Azure PaaS and SaaS deployments. Suggested preparation: compare data constraints, security, scalability, operations and integration, without assuming a particular deployment is already in place.

How would you integrate an AI service into existing business processes?

What to coverThe posting calls for integration with business processes and knowledge of Python, .NET, APIs and microservices. Suggested preparation: describe service boundaries, interfaces, failure handling and maintainability, and explain how the design would respect applicable internal policies.

How would you prepare and process data for an AI initiative?

What to coverData engineering principles, Databricks, Hadoop and Apache Spark appear in the posting, which also flags data limitations as a project risk. Suggested preparation: explain data quality checks, processing choices and what you would do if available data could not support the intended outcome.

How would you select and evaluate a non-GenAI machine-learning approach?

What to coverThe role spans supervised and unsupervised learning, NLP, computer vision and predictive analytics. Suggested preparation: choose a stated type of business problem, explain why a method fits, and define evaluation criteria and a baseline rather than assuming a specific project.

How would you address privacy, security and responsible AI in the architecture?

What to coverThe posting requires scalable, secure, maintainable solutions that meet data privacy requirements, industry standards and ethical guidelines. Suggested preparation: discuss data access, risk review and ongoing checks, while avoiding assumptions about which regulations or internal controls apply.

How would you evaluate AI platforms and troubleshoot weak solution performance?

What to coverThe architect is expected to assess tools and platforms, and the posting names Azure AI Foundry and Ollama while identifying algorithm performance as a risk. Suggested preparation: define comparison criteria and explain how you would isolate data, model and system causes before changing tools or architecture.

Behavioural practice

Tell me about a time you aligned technical work with a business objective.

What to coverThe role involves working with business units on ROI and KPIs. Suggested preparation: use a real situation, the actions you took to agree on measurable goals, and the result; do not claim an outcome you cannot substantiate.

Tell me about a time you coordinated people across different technical and business teams.

What to coverThe posting describes liaison work across research, data science, engineering, product and business stakeholders, including onsite and offshore teams. Suggested preparation: give the situation, how you clarified responsibilities or removed an obstacle, and the result.

Tell me about a time you handled competing project priorities or a delivery risk.

What to coverThe role includes managing multiple projects, milestones, risks and contingency plans. Suggested preparation: describe the situation, your prioritization or mitigation actions, and the result, including any trade-off you communicated.

Tell me about a time you worked with an external vendor or partner.

What to coverThe posting includes collaboration with external vendors and partners where appropriate. Suggested preparation: explain the situation, how you evaluated or coordinated the contribution, and the result without assuming a particular vendor arrangement.

Tell me about a time you raised a privacy or responsible-AI concern.

What to coverThe role calls for attention to data privacy, ethical guidelines and regulatory concerns. Suggested preparation: use a real situation, describe how you assessed and communicated the concern, and explain the resulting decision or change.

Sharpen your knowledge

Topics to prepare

GenAI architecture choices
The posting names RAG, LLMs, fine-tuning and text-to-SQL, and calls for experience with 3–4 GenAI solutions. As revision, take one hypothetical business problem and explain which approach you would choose, its trade-offs, and when a smaller language model might suffice.
Cloud and on-premises architecture
The role includes on-premises and cloud AI architecture, Azure and GCP, and Azure PaaS and SaaS deployments. As revision, sketch two deployment options for the same solution and explain scalability, security, maintainability and integration trade-offs.
ML methods and use cases
The listing names supervised and unsupervised learning, deep learning, NLP, computer vision and predictive analytics. As revision, map each method to a plausible business problem and explain how you would judge whether it fits.
Data engineering at scale
Data engineering principles and Databricks, Hadoop and Apache Spark appear in the posting. As revision, trace how data would move from ingestion through preparation to model use, identifying quality and scale risks.
Privacy, ethics and compliance
The architect must address data privacy, industry standards, ethical guidelines and applicable internal policies. As revision, review a hypothetical AI design for sensitive-data exposure, misuse and governance gaps, then propose mitigations without assuming a specific regulation.
AI delivery and business value
The role defines scope, milestones, ROI and KPIs, manages risks, and coordinates business and technical teams. As revision, draft a one-page plan for a hypothetical GenAI initiative with measurable KPIs, dependencies, risks and a contingency.

Skills to revise

Build a small RAG workflow
The posting explicitly names RAG, LangGraph and LangChain. For hands-on revision, use a small sample document set to implement a retrieval-and-answer flow, then test what happens when evidence is missing or irrelevant.
Integrate an AI service
Python, .NET, APIs and microservices are listed. For practice, define an API contract for a hypothetical AI feature and implement a small Python or .NET endpoint, including input validation and error handling.
Compare model-development options
The role covers development, training and optimization of large and small language models; Azure AI Foundry and Ollama are named tools. As a practice exercise, compare a small-model experiment with a proposed fine-tuning approach and record quality, deployment and maintenance trade-offs.
Plan an Azure deployment
The posting calls out Azure services and PaaS and SaaS deployment knowledge. For hands-on revision, produce a deployment diagram and checklist for a hypothetical AI application, covering access, data flow, monitoring and integration with an existing process.
Prepare data with Spark
The listing names Databricks and Apache Spark alongside data engineering. For practice, transform a small sample dataset with Spark, document quality checks, and explain what would need to change at larger scale.
Track an AI initiative
Technical project management and agile tools such as Jira and Trello are specified. As revision, turn the hypothetical initiative into milestones and tracked tasks, with owners, dependencies, risks and KPI review points.

Your preparation checklist

Map your evidence to the requirements
Prepare a one-page, factual map of relevant experience against the listing’s 10+ years overall, 5+ years in AI-related work, 4 years in GenAI/AI, and requested GenAI solutions. Identify gaps without inventing examples.
Build a GenAI project walkthrough
Prepare a concise walkthrough of a solution you actually worked on, covering the business problem, your architectural decisions, implementation, evaluation, results, and lessons learned. If available, outline additional RAG, LLM, fine-tuning, or text-to-SQL examples to address the listing’s 3–4 solution request.
Practice an architecture design exercise
Sketch a scalable, secure, maintainable AI solution and explain choices among on-premises deployment, Azure PaaS or SaaS, and GCP. Show how it would integrate with APIs, microservices, data pipelines, and an existing business process.
Prepare to defend model and tool choices
Create a short decision table comparing RAG, fine-tuning, and other model approaches for a business use case. Be ready to explain relevant choices involving LangGraph, LangChain, Azure AI Foundry, Databricks, or Ollama without claiming hands-on use you cannot substantiate.
Outline delivery and measurement
Draft a sample scope, milestones, risks, contingency actions, KPIs, and approach to estimating ROI for an in-house GenAI platform initiative. Include how you would prioritize work and coordinate business stakeholders and onsite and offshore teams.
Review responsible-AI controls
Prepare concrete talking points on data privacy, access controls, model evaluation, ethical risks, and regulatory considerations. Distinguish controls you have implemented from those you would propose; the listing also mentions compliance with Arrow’s internal policies.
Write targeted questions for the employer
Ask which organization owns the AI work and applicable internal policies, what stage the in-house GenAI platform has reached, how success and ROI are measured, and how work is divided across teams. Confirm whether the listed BangalorePune location means Bangalore, Pune, or either.

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…