New opportunity

SENIOR AI ENGINEER

A4I Lab (IIIT-B × Microsoft) · Bangalore, India

About this role

The A4I Lab (a partnership between Microsoft and IIIT-B) is seeking a Senior AI Engineer to design and scale agentic systems for social impact in sectors like education, healthcare, and agriculture. This role focuses on building production-ready AI platforms and multi-agent workflows that are reliable, scalable, and suitable for real-world operating conditions.

Responsibilities

Lead the architecture of end-to-end agentic AI systems from prototype to production.
Build integrated AI platforms combining LLMs, retrieval systems, and multi-modal inputs (speech, documents, vision).
Define architectural standards for deployment, monitoring, versioning, and safe rollout.
Establish sound engineering and MLOps practices for consistency and traceability.
Provide technical leadership, mentor team members, and contribute to open-source Digital Public Goods.

Requirements

6 to 8 years of experience in software systems, with at least 2 years focused on production AI or ML applications.
Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, Mathematics, or equivalent.
Hands-on experience with agent orchestration frameworks (e.g., LangGraph, LlamaIndex, n8n).
Strong understanding of RAG architectures, context management, and evaluation.
Proficiency in Python for production environments.
Experience with cloud platforms (Azure preferred) and containerization.

Preferred Qualifications

Experience adapting open models (Llama, Mistral, Phi) using PEFT or LoRA.
Exposure to VLMs, speech models, and multilingual pipelines.
Familiarity with AI evaluation frameworks like RAGAS or LangSmith.
Prior work in social impact or resource-constrained settings.
Contributions to open-source AI projects.

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

Skills for this role

PythonAgentic SystemsLangGraphLlamaIndexn8nLLMRAGVector RetrievalCloud EngineeringAzureContainerizationCI/CDMLOpsPEFTLoRAInstruction TuningVLMsSpeech ModelsOllamavLLMHugging FaceRAGASLangSmith

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

Production-Grade Agentic Architecture
The primary focus is designing and scaling end-to-end agentic systems that transition from prototype to reliable, production-ready platforms. You must demonstrate the ability to structure multi-step workflows, tool use, and memory management using frameworks like LangGraph or LlamaIndex while ensuring system reliability and maintainability.
MLOps and Platform Maturity
You are expected to establish rigorous engineering standards for deployment, versioning, monitoring, and safe rollout. This involves implementing MLOps practices that ensure traceability and consistency, moving beyond experimental code to robust, production-grade AI infrastructure.
Social Impact and Resource-Constrained Design
The role requires building systems for real-world operating conditions, including multilingual use cases and resource-constrained environments. You must balance performance, cost, and resilience to ensure that AI solutions are accessible and effective for diverse stakeholders in sectors like healthcare and education.
Technical Leadership and Open-Source Contribution
As a senior member, you will guide architectural decisions, mentor team members, and contribute to the development of Digital Public Goods. You must be able to communicate complex technical trade-offs to cross-functional stakeholders and foster a culture of reusable, open-source AI development.

Likely challenges to think through

Balancing Innovation with Reliability
The challenge lies in applying cutting-edge agentic frameworks to high-stakes social impact sectors where system failure has real-world consequences. Prepare by articulating how you implement guardrails, evaluation frameworks like RAGAS, and robust testing to ensure system reliability in unpredictable environments.
Managing Multimodal and Multilingual Complexity
Integrating speech, vision, and document processing into a cohesive pipeline while supporting multilingual requirements introduces significant architectural overhead. Prepare by discussing your experience with model integration patterns and how you handle data quality and context management across diverse input types.
Navigating Multi-Stakeholder Constraints
Working between academia, Microsoft, and non-profit partners requires aligning technical vision with diverse organizational goals. Prepare by highlighting your experience in managing expectations and translating technical requirements into solutions that satisfy both engineering rigor and social impact objectives.

Practice your answers

Technical practice

Designing Reliable Agentic Workflows

What to coverExplain how you ensure reliability and error handling in multi-agent systems using frameworks like LangGraph. Discuss strategies for managing state, handling tool-use failures, and implementing retry logic in production environments.

Optimizing RAG for Domain-Specific Accuracy

What to coverDescribe your approach to improving retrieval quality in RAG architectures. Address techniques such as hybrid search, re-ranking, and context management to ensure grounding and minimize hallucinations in specialized domains.

Production MLOps for LLM Applications

What to coverDetail the MLOps lifecycle for agentic systems, specifically regarding versioning, monitoring, and safe rollout. Explain how you would implement observability to track system performance and drift in a production setting.

Adapting Open Models for Specific Tasks

What to coverDiscuss the trade-offs between using proprietary APIs versus hosting open models like Llama or Mistral. Explain how you would apply PEFT or LoRA to fine-tune these models for domain-specific requirements while maintaining efficiency.

Evaluating AI System Performance

What to coverHow do you define and measure success for an agentic system? Discuss the use of evaluation frameworks like RAGAS or LangSmith and how you create custom benchmarks to validate system quality.

Handling Multi-modal Inputs in Production

What to coverExplain the architectural challenges of integrating speech, vision, and document processing into a unified agentic pipeline. Focus on data synchronization, latency management, and model orchestration.

Scaling Systems for Resource-Constrained Environments

What to coverDescribe your strategy for deploying AI solutions in environments with limited infrastructure or multilingual requirements. Suggest how you balance model size, inference cost, and system responsiveness.

Python Engineering Standards for AI

What to coverWhat practices do you follow to ensure AI codebases remain maintainable and testable? Discuss your approach to modular design, dependency management, and CI/CD pipelines for AI-heavy projects.

Behavioural practice

Leading Technical Architecture Decisions

What to coverDescribe a situation where you had to lead the architectural design of a complex system. Explain how you balanced competing technical requirements and communicated the trade-offs to stakeholders.

Mentoring and Technical Leadership

What to coverProvide an example of how you have mentored team members or guided a team through a difficult technical transition. Focus on how you fostered a culture of sound engineering and knowledge sharing.

Navigating Multi-stakeholder Environments

What to coverShare an experience working with diverse stakeholders, such as non-technical partners or academic researchers. How did you align their expectations with the technical realities of the project?

Contributing to Open Source or Public Goods

What to coverDiscuss your experience contributing to open-source projects or building reusable assets. Explain how you ensure your work is documented and accessible for others to build upon.

Managing Project Scope for Social Impact

What to coverDescribe a time you had to adapt a project to meet the needs of a resource-constrained setting. How did you prioritize features to ensure the system remained usable and impactful?

Sharpen your knowledge

Topics to prepare

Agentic System Architecture
Focus on designing multi-agent workflows that handle state management, tool invocation, and error recovery. Exercise: Draft a high-level architecture diagram for a multi-agent system that processes multilingual documents, explaining how you would handle agent hand-offs and state persistence in a production environment.
Production RAG and Evaluation
Understand the lifecycle of RAG systems from retrieval optimization to output evaluation. Exercise: Define a strategy for evaluating a RAG pipeline using metrics like faithfulness and answer relevance, and explain how you would implement a feedback loop to improve retrieval quality over time.
MLOps for AI Platforms
Focus on the operational requirements for deploying LLM-based applications. Exercise: Outline a CI/CD pipeline for an AI service that includes automated model versioning, drift monitoring, and a safe rollout strategy to ensure system reliability in production.

Skills to revise

Agent Orchestration Frameworks
Proficiency in tools like LangGraph or LlamaIndex is required. Exercise: Build a simple multi-step agent workflow using LangGraph that performs a web search, summarizes the findings, and formats the output, ensuring the code is modular and testable.
Model Adaptation (PEFT/LoRA)
Focus on fine-tuning open models for specific domains. Exercise: Create a script using the PEFT library to apply LoRA to a small open-source model (e.g., Phi or Llama) for a specific classification task, documenting the memory and performance trade-offs.
Python for Production Systems
Demonstrate clean, maintainable, and scalable coding practices. Exercise: Refactor a prototype script into a production-ready Python module, incorporating type hinting, comprehensive logging, and unit tests to ensure it meets enterprise-grade standards.

Your preparation checklist

Prepare an end-to-end agentic system case study
Document a project where you moved an AI application from prototype to production. Focus on your architectural choices regarding agent orchestration (e.g., LangGraph or LlamaIndex), handling multi-modal inputs, and ensuring system reliability in real-world conditions.
Demonstrate MLOps and production-readiness expertise
Prepare to discuss your approach to CI/CD for AI, model versioning, and monitoring. Be ready to explain how you implement evaluation frameworks like RAGAS or LangSmith to maintain performance and traceability in production environments.
Showcase experience with open-source and domain-specific tuning
Gather evidence of your work with open models like Llama or Mistral. Specifically, prepare to talk about your experience with PEFT, LoRA, or instruction tuning to adapt models for domain-specific tasks or resource-constrained settings.
Formulate questions on Digital Public Goods and social impact
Prepare thoughtful questions about the lab's strategy for scaling Digital Public Goods. Ask how they balance technical performance with the unique requirements of sectors like education or healthcare, and how they manage multi-stakeholder collaboration.
Review cloud and infrastructure design patterns
Be prepared to discuss your experience with Azure or similar cloud platforms, specifically regarding containerization and deploying scalable AI services. Highlight your ability to design systems that remain performant and cost-effective under real-world constraints.

Availability last checked: 18 Oct 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…