New opportunity

Engineering Lead - Applied AI

Barclays · Pune, India

About this role

Barclays is seeking an Engineering Lead - Applied AI in Pune to advance its data technologies through AI and build robust, scalable, secure applications. The role combines applied AI engineering with software delivery, technical leadership and responsible use of AI.

Responsibilities

Design, develop and improve software that provides business, platform and technology capabilities. Deliver maintainable, performance-optimized solutions using appropriate languages, frameworks and tools.
Work with product managers, designers and engineers to define requirements, plan solutions and integrate them with business objectives. Participate in code reviews, share knowledge, apply secure coding practices and implement unit testing.
Provide leadership in software architecture, project planning and execution, mentoring and cross-team collaboration. At Vice President level, contribute to strategy and requirements, advise senior stakeholders, manage risks and controls, and drive improvements. Depending on whether the role has people-management or individual-contributor responsibilities, lead and develop a team or guide technical direction and coach specialists.

Required or assessed capabilities:

Applied AI experience with tools and techniques including Claude CLI, OpenAI SDK, Google ADK, RAG, LangChain, Langfuse, Bedrock, guardrails, vector databases and advanced prompting.
Working knowledge of AWS and associated technologies, including Python, IAM, S3, Lambda, Step Functions, CloudWatch, Glue, Athena, Lake Formation, RDS, Postgres, ECS, EKS and CloudFormation.
Technical leadership exposure across software architecture, line management, mentoring, cross-team collaboration, and project planning and execution. Strong client-facing and stakeholder communication is required. Assessment may also cover risk and controls, transformation, business acumen and strategic thinking.

Desirable: Experience with Hadoop, Impala or Hive; cloud data technologies such as Snowflake, Databricks, Starburst and StarRocks; and data governance and controls.

This is a permanent, full-time role based in Pune. Barclays describes a hybrid approach with onsite anchor days; the role-specific pattern should be discussed with the hiring manager. Listed benefits include holiday allowance, life assurance, private medical care and a pension contribution.

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

Skills for this role

Applied AIClaude CLIOpenAI SDKGoogle ADKRAGLangChainLangfuseAmazon BedrockAI guardrailsVector databasesPrompt engineeringAWSPythonIAMAmazon S3AWS LambdaAWS Step FunctionsAmazon CloudWatchAWS GlueAmazon AthenaAWS Lake FormationAmazon RDSPostgreSQLAmazon ECSAmazon EKSAWS CloudFormationSoftware architectureTechnical leadershipLine managementMentoring project teams

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

Deliver applied AI that is robust and responsible
The posting calls for advancing data technologies through AI and building scalable, secure applications. Applied AI experience, including RAG, guardrails and vector databases, is among the capabilities that may be assessed.
Own software quality from design through delivery
Accountabilities include maintainable, performance-optimized software, code reviews, secure coding and unit testing. Prepare to explain architectural decisions and how quality was maintained during delivery.
Connect AI solutions to the AWS data landscape
The role calls for working knowledge of AWS, Python and listed data, compute, security and deployment services. Focus on how these components support a solution rather than reciting the service list.
Lead across teams and advise stakeholders
Technical leadership, planning, mentoring and client-facing communication are stated capabilities. Vice President expectations also include shaping strategy and requirements, advising senior stakeholders, and managing risks and controls; the posting describes both people-leader and individual-contributor paths.

Likely challenges to think through

Balancing AI usefulness with control
Preparation hypothesis: improving an AI application's utility may create tension with protecting sensitive data and managing risk. Prepare an example of how you would weigh output quality against guardrails, secure coding and governance, without assuming a particular Barclays policy.
Choosing an architecture without overcomplicating it
Preparation hypothesis: the breadth of AI and AWS technologies could make it difficult to choose a maintainable design that also performs and scales. Prepare to compare a few plausible component choices and explain what you would test before committing.
Leading when the reporting model is unclear
Preparation hypothesis: influence may be needed across teams whether this position manages people or guides technical direction as an individual contributor. Prepare distinct examples of mentoring, resolving cross-team trade-offs and communicating decisions to senior stakeholders.

Practice your answers

Technical practice

How would you design a retrieval-augmented generation system for a business application?

What to coverSuggested preparation: Explain document ingestion, retrieval with a vector database, grounding responses in retrieved material, and handling poor or missing matches. Discuss trade-offs among answer quality, latency, scalability and access to sensitive data.

How would you choose among AI SDKs and orchestration tools for an application?

What to coverSuggested preparation: Compare where an SDK or service such as OpenAI SDK or Bedrock might suffice and where a framework such as LangChain or Google ADK could help. Explain how you would assess complexity, maintainability and integration needs rather than assuming one tool fits every use case.

How would you combine advanced prompting with AI guardrails?

What to coverSuggested preparation: Describe how you would specify the task, constrain outputs and test prompts against ambiguous or unsafe inputs. Explain where application-level validation and controls remain necessary even when prompts are well designed.

How would you evaluate and troubleshoot an applied AI feature?

What to coverSuggested preparation: Describe checks for answer relevance, groundedness, failures, latency and changes after a prompt or retrieval update. The posting lists Langfuse; discuss how tracing could help isolate a problem without assuming Barclays uses a particular evaluation process.

How would you design an AWS data flow to support an AI application?

What to coverSuggested preparation: Use relevant listed services, such as S3, Glue, Athena and Step Functions, to explain ingestion, preparation and orchestration. Address data quality, failure handling and when a different storage or query approach would be preferable.

How would you choose an AWS runtime and deployment approach?

What to coverSuggested preparation: Compare Lambda with ECS or EKS for workload duration, scaling, operational effort and application requirements. Explain how CloudFormation and CloudWatch could support repeatable deployment and operational visibility.

How would you protect sensitive data across an AI application and its data stores?

What to coverSuggested preparation: Discuss least-privilege IAM access, appropriate data controls and secure handling across services such as S3, Lake Formation and RDS or PostgreSQL. Explain how you would review risks at system boundaries rather than treating model guardrails as the only protection.

How would you maintain code quality while delivering applied AI software?

What to coverSuggested preparation: Cover code-review criteria, unit tests for deterministic components, and testing around external AI responses. Explain how you would balance readable, maintainable code with performance and delivery needs.

How would you investigate a slow or unreliable AI-backed application?

What to coverSuggested preparation: Break the request path into retrieval, model calls, application code and infrastructure, then describe what you would measure to locate the issue. Compare improvements against their effects on response quality, cost and operational complexity.

When might you consider a different data platform for an AI workload?

What to coverSuggested preparation: The posting marks Hadoop, Impala, Hive and cloud data technologies such as Snowflake and Databricks as desirable, not mandatory. Compare platforms at a high level using data scale, query patterns, governance and integration needs, without assuming a particular Barclays platform.

Behavioural practice

Tell me about a time you aligned stakeholders on a technical direction.

What to coverSuggested practice question: Give the situation, the competing business and engineering priorities, your actions to explain options to stakeholders, and the result. Include how you handled disagreement or changed the plan in response to feedback.

Describe a time you developed others while maintaining delivery standards.

What to coverSuggested practice question: Use a real mentoring or team-leadership example, stating the situation, your coaching actions and the result. Make clear whether you had people-management responsibility or were guiding specialists through technical leadership.

Tell me about a risk or control concern you identified during delivery.

What to coverSuggested practice question: Explain the situation, how you assessed and communicated the concern, what action you took or escalated, and the result. Focus on your actual decisions and evidence rather than claiming a control was effective without support.

Describe a cross-functional project where requirements changed.

What to coverSuggested practice question: Explain the situation, how you worked with product, design or engineering partners to revise the plan, and the result. Address the trade-off between delivery timing and the quality or maintainability of the solution.

Tell me about a change you led to improve an engineering approach.

What to coverSuggested practice question: Describe the situation, your rationale, how you gained support and executed the change, and the result. Connect the example to strategic priorities or continuous improvement without implying a specific Barclays initiative.

Sharpen your knowledge

Topics to prepare

Applied AI and retrieval architecture
The posting names RAG, vector databases, advanced prompting, LangChain, Google ADK and model SDKs among skills that may be assessed. As a revision exercise, sketch a hypothetical retrieval-based application and explain how documents become searchable, how context reaches a model, and when orchestration adds value.
AWS application and data architecture
The posting calls for working knowledge of Python and AWS services including IAM, S3, Lambda, Step Functions, Glue, Athena, Lake Formation, RDS, Postgres, ECS, EKS and CloudFormation. Draw a small application architecture using relevant listed services, then justify service choices, access boundaries and scaling trade-offs.
Responsible AI, risk and controls
The role emphasizes responsible AI and may assess guardrails, risk and controls; data governance is desirable. For a hypothetical AI response containing sensitive data, identify failure points and explain proposed access, output and monitoring controls. Treat the design as a preparation exercise, not a stated Barclays implementation.
Software quality and security
The accountabilities explicitly include maintainable, scalable, performance-optimized code, secure coding, code reviews and unit testing. Review a short design on paper and explain which risks you would test, which changes you would request in review, and how you would assess performance.
Technical leadership and stakeholder decisions
The posting specifies architecture, mentoring, cross-team delivery, planning and strong client-facing communication. Its Vice President expectations cover strategy, senior stakeholders and risk, with different paths for people leaders and individual contributors. Rehearse a concise decision brief comparing two technical options, their business impact, delivery risks and a recommendation.

Skills to revise

Prototype a small RAG flow
Build a toy retrieval-and-answer flow using a vector database and one model integration named in the posting, such as the OpenAI SDK or Bedrock. Test a few questions for retrieval relevance and unsupported answers. This practices the listed applied AI capabilities; the prototype is a suggested exercise, not a specified deliverable.
Implement a Python AWS workflow
As a practice exercise, write Python that accepts a small sample input, stores it in S3 and invokes Lambda; sketch where Step Functions and CloudWatch would fit. Explain the IAM permissions you would limit. Python and these AWS services appear in the working-knowledge list.
Add guardrails and observability
For the toy AI flow, add a check that rejects an unsafe or unsupported answer and inspect a trace using Langfuse if available. Record what the check catches and misses. Guardrails and Langfuse are named applied AI skills, while this particular implementation is only a revision suggestion.
Review, secure and unit-test code
Take a small Python function, write unit tests for normal and failure cases, then perform a code review focused on readability, sensitive-data handling and performance. The posting explicitly calls for unit testing, secure coding, code reviews and maintainable solutions.
Present an architecture decision
Prepare a brief walkthrough of a hypothetical AI application for both engineers and a senior stakeholder: state the requirement, compare options, explain controls and propose a delivery sequence. This rehearses the posting’s cross-functional planning, technical leadership and stakeholder communication expectations.

Your preparation checklist

Prepare an applied AI project walkthrough
The role lists RAG, model SDKs, vector databases, guardrails and observability among skills that may be assessed. Prepare one project walkthrough showing the problem, your design choices, evaluation approach, safeguards and results; identify which listed tools you actually used.
Sketch a secure AWS design
As a practice exercise, draw a scalable applied AI service using relevant listed technologies such as Python, IAM, S3, Lambda or ECS, and CloudWatch. Be ready to explain access control, data flow, deployment choices and trade-offs without assuming this is Barclays’ architecture.
Collect software-delivery evidence
Prepare a concise example of how you improved maintainability or performance, with a code-review decision, unit-testing approach and secure-coding measure. These map to the role’s stated engineering accountabilities.
Map your AWS and data-technology experience
Make a one-page evidence matrix for the listed AWS services and Python, distinguishing hands-on work from working knowledge. Put Hadoop, Hive, Impala and the named cloud data platforms in a separate desirable-skills section.
Rehearse a leadership and delivery example
Prepare an example covering architecture direction, planning, cross-team coordination, mentoring and an outcome. Separate any people-management experience from technical guidance, since the posting describes both possible leadership paths.
Prepare a risk and stakeholder discussion
The posting says risk, controls, strategic thinking and stakeholder communication may be assessed. Outline a decision you can explain in terms of options, business impact, controls, escalation and how you communicated with stakeholders.
Write questions for the hiring manager
Ask whether this position has people-management or individual-contributor responsibilities, which applied AI and data challenges take priority, how responsible AI and data governance are handled, and what onsite anchor-day pattern applies in Pune.

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…