New opportunity

AI Engineer

Shuru Technologies Private Limited · Gurgaon, India

About this role

Shuru Technologies is an AI, product, and technology engineering firm focused on building and scaling digital products for startups and enterprises. We are seeking a hands-on AI Engineer to transform ambiguous requirements into production-grade AI systems. You will be responsible for building AI agents, intelligent workflows, and automations, managing the full lifecycle from prototype to production. Responsibilities include developing memory, context, and state management layers, creating frameworks for prompts, tools, and guardrails, and implementing robust evaluation and observability systems for AI performance. You will also build secure APIs and integrate agents with enterprise systems and event streams. Requirements: Minimum 5 years of experience in Python and backend development, including FastAPI, REST APIs, and microservices. At least 2 years of experience building agentic or LLM applications using frameworks like LangGraph or LangChain. Proficiency in memory/context engineering, workflow testing, and AI observability is essential. Experience with MongoDB or document databases and cloud-native environments like AWS is required. Preferred qualifications include experience with Docker, Kubernetes, Kafka, AWS Bedrock, vector databases, MCP, knowledge graphs, and human-in-the-loop workflows, as well as experience in regulated domains like healthcare. Benefits include working on international projects, team-building events, and a collaborative growth-oriented environment.

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

Skills for this role

PythonFastAPIREST APIsMicroservicesLangGraphLangChainMemory and context engineeringAgent evaluationWorkflow testingAI observabilityMongoDBAWSDockerKubernetesKafkaAWS BedrockVector databasesMCPKnowledge graphsHuman-in-the-loop workflows

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 AI Agent Lifecycle Management
The role requires moving beyond prototypes to deploy reliable AI agents. You must demonstrate expertise in building memory, context, and state management layers, while ensuring these systems are production-ready through robust evaluation and regression testing frameworks.
Scalable Backend and API Integration
As an AI Engineer at Shuru, you must bridge the gap between AI models and enterprise systems. This involves building secure, high-performance APIs using FastAPI and microservices, and integrating agents with event streams, vector databases, and existing enterprise infrastructure.
AI Observability and Quality Assurance
Maintaining AI performance is a core priority. You are expected to implement comprehensive observability systems to track latency, token usage, error rates, and output quality, ensuring that agentic workflows remain stable and predictable in production environments.

Likely challenges to think through

Managing Ambiguity in AI Requirements
The role involves transforming ambiguous business requirements into technical AI solutions. Prepare by practicing how you translate high-level business goals into specific agentic workflows, guardrails, and tool definitions, ensuring you can define success metrics when the path forward is not clearly mapped.
Balancing Agent Autonomy with Guardrails
Building intelligent workflows requires a trade-off between agent flexibility and system safety. You will face the challenge of implementing effective guardrails and human-in-the-loop workflows that prevent errors without stifling the agent's ability to solve complex, multi-step tasks.

Practice your answers

Technical practice

Designing Agentic Memory and State Management

What to coverExplain your approach to implementing long-term memory and state persistence in LLM agents. Discuss the trade-offs between using vector databases versus traditional document stores like MongoDB for maintaining context across multi-turn interactions.

Implementing AI Observability and Evaluation

What to coverDescribe how you would build an observability pipeline to track latency, token usage, and hallucination rates in production. Suggest specific metrics or frameworks you would use to perform regression testing on agentic workflows.

Building Robust Agentic Workflows with LangGraph

What to coverDiscuss the advantages of using state-machine based frameworks like LangGraph over simple sequential chains. Provide an example of how you handle complex branching logic and error recovery within an agentic loop.

Integrating LLMs with Enterprise Event Streams

What to coverExplain how you would integrate an AI agent with a Kafka event stream to process real-time data. Address how you ensure system reliability and handle asynchronous communication between the agent and external APIs.

Prompt Engineering and Guardrail Frameworks

What to coverHow do you structure prompts and tool definitions to ensure consistent agent behavior? Describe your methodology for implementing guardrails to prevent prompt injection and ensure output validation in production environments.

Scaling Microservices with FastAPI and Kubernetes

What to coverDescribe your process for containerizing and deploying Python-based AI services using Docker and Kubernetes. Focus on how you manage resource allocation and scaling for compute-intensive LLM inference tasks.

Human-in-the-Loop (HITL) Implementation

What to coverExplain how you design workflows that require human intervention for critical decision-making. Discuss how you manage state persistence while waiting for human feedback within an automated agentic process.

Knowledge Graphs in RAG Architectures

What to coverWhen would you choose to augment a standard vector-based RAG system with a knowledge graph? Explain how you would manage the synchronization between structured knowledge and unstructured vector embeddings.

Behavioural practice

Turning Ambiguous Requirements into Production Systems

What to coverDescribe a situation where you were given an ambiguous business problem and had to define the technical scope for an AI solution. Focus on how you gathered requirements and validated your prototype before moving to production.

Handling Production Failures in AI Systems

What to coverProvide an example of a time an AI system you built failed in production. Explain how you diagnosed the issue, the steps you took to mitigate the impact, and how you improved your observability to prevent recurrence.

Collaborating with Cross-functional Stakeholders

What to coverDescribe a time you had to explain the limitations or capabilities of an AI model to non-technical stakeholders. How did you manage their expectations while ensuring the project remained technically feasible?

Managing Technical Debt in Rapidly Evolving AI Projects

What to coverAI frameworks evolve quickly; describe a time you had to refactor a significant portion of an application due to changing technology or requirements. How did you balance the need for speed with the need for maintainable, production-grade code?

Sharpen your knowledge

Topics to prepare

Agentic State and Memory Management
Focus on designing persistent state layers for AI agents. Exercise: Architect a system that maintains conversation history and user context across multiple sessions using a document store like MongoDB. This is relevant because the role requires building memory and context management layers for production-grade AI systems.
AI Observability and Evaluation Frameworks
Study methodologies for tracking latency, token usage, and output quality in LLM pipelines. Exercise: Create a regression test suite for an agentic workflow that measures success rates against a set of ground-truth inputs. This is essential for the role's requirement to implement robust evaluation and observability systems.
Human-in-the-Loop (HITL) Workflow Design
Understand patterns for integrating human approval steps into automated agent workflows. Exercise: Design a state machine using LangGraph that pauses execution for human intervention when a confidence threshold is not met. This is a preferred qualification for building reliable enterprise-grade automations.

Skills to revise

LangGraph and LangChain Implementation
Practice building complex, multi-step agentic workflows. Exercise: Build a multi-agent system where one agent handles tool selection and another handles response generation, orchestrated via LangGraph. This directly addresses the requirement for 2+ years of experience in agentic application development.
FastAPI and Microservices Integration
Develop secure, scalable backend services. Exercise: Build a FastAPI service that exposes an agentic workflow as a REST API, including authentication and asynchronous processing for long-running tasks. This is critical for the role's focus on building secure APIs and integrating agents with enterprise systems.
Vector Database and Retrieval Integration
Master the retrieval-augmented generation (RAG) pipeline. Exercise: Implement a document retrieval system that fetches context from a vector database based on user queries to inform agent responses. This is relevant to the requirement for integrating agents with vector databases and enterprise data.

Your preparation checklist

Prepare a technical walkthrough of an agentic workflow
Select a project where you implemented an AI agent using LangGraph or LangChain. Prepare to explain how you handled state management, memory, and context retrieval, specifically focusing on how you moved the system from a prototype to a production-grade environment.
Develop a case study on AI observability and evaluation
Document a specific instance where you implemented observability tools to track latency, token usage, and error rates. Explain the regression testing framework you used to evaluate agent performance and how you ensured reliability in a production setting.
Review backend architecture for microservices and APIs
Be prepared to discuss your experience building secure REST APIs with FastAPI. Specifically, prepare examples of how you integrated these services with event streams (like Kafka) or document databases (like MongoDB) within a cloud-native AWS environment.
Formulate questions regarding their AI-native engineering culture
Prepare thoughtful questions about how Shuru Technologies manages the ambiguity of client requirements. Ask about their specific approach to human-in-the-loop workflows or how they balance rapid experimentation with the need for robust, production-grade guardrails.

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…