Job Description

About KPMG in India

KPMG entities in India are professional services firm(s). These Indian member firms are affiliated with KPMG International Limited. KPMG was established in India in August 1993. Our professionals leverage the global network of firms, and are conversant with local laws, regulations, markets and competition. KPMG has offices across India in Ahmedabad, Bengaluru, Chandigarh, Chennai, Gurugram, Jaipur, Hyderabad, Jaipur, Kochi, Kolkata, Mumbai, Noida, Pune, Vadodara and Vijayawada.

KPMG entities in India offer services to national and international clients in India across sectors. We strive to provide rapid, performance-based, industry-focused and technology-enabled services, which reflect a shared knowledge of global and local industries and our experience of the Indian business environment.

Overview

This role demand strong GenAI experience, emerging mastery in Agentic AI Systems, and a good foundation in classical ML.

You will design and build intelligent, tool-using agents, multi-agent systems, RAG pipelines, and LLM-based applications leveraging the LangChain , LangGraph ecosystem, LangSmith for evaluation.

________________________________________

Key Responsibilities

GenAI / LLM Application Development
Build GenAI applications using:

oLangChain, LangGraph

Implement RAG architectures with:

oRetrieval, reranking, chunking, memory strategies

oVector DBs (faiss, aisearch, opensearch, PG vector etc).

Design prompt-engineering strategies:

oInstruction-following

oReAct (Reasoning + Acting)

oChain-of-thought structuring

oSelf-reflection and planning loops

Evaluation Strategy

oImplement evaluation frameworks for Classical ML and GenAI systems, covering statistical validation, reliability, and robustness.

oAssess LLM outputs, RAG pipelines, and agent workflows for grounding quality, relevance, and retrieval accuracy (e.g., recall@k, precision@k).

oUse LangSmith for tracing, automated evaluations, regression testing, and continuous system level quality monitoring

Agentic System Architecture
Build agentic workflows:

oTool-calling agents

oPlanner–executor systems

oMulti-agent communication systems

oHierarchical agent architectures

oDeep Agents

Integrate memory systems:

oepisodic memory

osemantic memory

ovector-based long-term knowledge

Implement evaluation frameworks for agentic systems using LangSmith.
Model Context Protocol (MCP) & Tooling
Implement MCP servers for external tool connectivity.
Build tools that allow agents to interact with:

oAPIs

oCode execution environments

oKnowledge bases

oCompany applications

Classical ML (Foundational DS Skills)
Apply ML models to structured/unstructured data.
Conduct feature engineering, model selection, hyperparameter tuning.
Build interpretable models where required.
Engineering & Integration
Collaborate with backend engineering teams to seamlessly integrate agentic and GenAI systems into production applications.
Implement observability, tracing, and monitoring for GenAI workflows using LangSmith to ensure reliability and system‑level transparency.
Cloud ML-Ops & Quality
ML Modelling, data drift, concept drift, model quality monitoring.
Hands‑on experience across AWS/ Azure/ Databricks, with flexibility to work on any cloud platform.
Adhere to stringent quality assurance and documentation standards using version control and code repositories (e.g., Git, GitHub, Markdown)

________________________________________

Required Skills & Experience

5–10 years total experience, with 2–4+ years hands-on GenAI.
Hands-on expertise with:

oLangChain, LangGraph

oLangSmith (tracing, metrics, evaluations)

oMCP tooling and agent tool integration

oReAct, Tree of Thoughts, multi-agent orchestration

oRAG patterns and vector databases

Strong coding expertise in Python.
Classical ML foundations (tree models, regression, etc.).
Experience working with LLM APIs and/or open-source LLMs.
Experience building and debugging production-quality GenAI pipelines.
Aws/azure
GIT Ops
Prior experience building complex multi-agent systems for real-world applications.
Knowledge of multi-modal LLMs (vision, speech, code).
Familiarity with structured evaluation of LLM systems (hallucination tests, safety assessments etc ).
Experience in enterprise-grade LLM deployments.

Equal employment opportunity information

KPMG India has a policy of providing equal opportunity for all applicants and employees regardless of their color, caste, religion, age, sex/gender, national origin, citizenship, sexual orientation, gender identity or expression, disability or other legally protected status. KPMG India values diversity and we request you to submit the details below to support us in our endeavor for diversity. Providing the below information is voluntary and refusal to submit such information will not be prejudicial to you.

Qualifications

BTech

KPMG India

New Jobs

Manager - DevOps