AI Automation Engineer Associate

JPMorgan Chase & Co.; Bengaluru; On-site; INR 3,500,000 - 5,500,000; Full time

Job summary

JPMorgan Chase & Co. is seeking a Senior Associate in Operations Automation in Bengaluru to design, build, deploy, and operate production-grade AI assistants that automate high-volume knowledge work.

You will create intelligent document processing pipelines, convert unstructured inputs to structured data, and monitor production quality with Responsible AI controls. You will define success metrics, develop end-to-end solutions, and collaborate with stakeholders to ensure reliability, scalability,

Qualifications

• At least 5 years of hands-on experience in applied AI, ML, or AI-powered automation.
• Strong Python skills including asynchronous programming.
• Experience with LLM techniques, prompt engineering and structured outputs.
• Proficiency with graph-based orchestration (LangGraph or equivalent).
• Define and validate schemas using Pydantic and data processing with pandas.
• Experience with OCR and document digitization workflows (PDF text extraction).
• Integrate external tools via APIs with robust validation and error handling.
• Sound software engineering practices: testing, CI/CD, secure development, production readiness.
• Git and code review workflows.
• Define objective success metrics and communicate across stakeholders.
• Root-cause analysis and measurable remediation.

Responsibilities

• Design production-grade AI assistants that combine LLMs, multimodal models, rules, retrieval, and tool integrations to automate workflows.
• Engineer multi-stage AI workflows with ingestion, model execution, tool invocation, validation, and post-processing.
• Apply architecture techniques across structured outputs, retrieval-augmented generation, and multimodal document processing.
• Define output schemas (e.g., Pydantic), prompts, deterministic post-processing, and exception handling.
• Build document digitization and OCR pipelines, including PDF text extraction and noisy scan handling.
• Establish success metrics and monitor quality, drift, latency, and cost in production.
• Write secure, maintainable production code and reusable components across use cases.
• Troubleshoot production issues across model, retrieval, tooling, and post-processing layers.
• Embed Responsible AI governance into delivery for auditability and traceability.
• Produce health views and KPIs for adoption, accuracy, and exception rates for stakeholders.
• Facilitate requirements elicitation with operations teams to translate needs into controls.

Skills

Applied AI experience

Python

LLM techniques

LangGraph

Pydantic

Pandas

OCR / document digitization

APIs

CI/CD

Git

Metrics

Root-cause analysis

Education

Master’s degree in AI/ML or related

Tools

AWS Textract

Job description

Make your mark by building reliable AI automation that improves accuracy, reduces manual effort, and grows your engineering impact.

As a Senior Associate in Operations Automation within the Commercial and Investment Bank, you design, build, deploy, and operate production-grade AI assistants that automate high-volume knowledge work. You focus on intelligent document processing by transforming unstructured inputs (such as PDFs, spreadsheets, emails, and images) into structured, validated data for downstream business systems. You define success metrics and continuously improve quality through offline evaluation and production monitoring while embedding Responsible AI controls and troubleshooting issues across the full solution stack.

Job Responsibilities:

• Design production-grade AI assistants that combine language models, multimodal models, business rules, retrieval, and tool integrations to automate operational workflows.
• Engineer multi-stage workflows using directed-graph or graph-based orchestration patterns, including ingestion, model execution, tool invocation, validation, and post-processing.
• Apply fit-for-purpose techniques across structured model outputs, retrieval-augmented generation, agentic patterns, and multimodal document processing.
• Define output schemas (for example, Pydantic), prompts, deterministic post-processing rules, validation logic, and exception handling aligned to domain needs.
• Build document digitization and Optical Character Recognition pipelines, including PDF text extraction, OCR processing, normalization, and handling of noisy or low-quality scans.
• Establish success metrics and improve accuracy and reliability through offline evaluation (golden datasets, regression tests, error analysis) and production monitoring (quality, drift, latency, cost).
• Write secure, maintainable, well-tested production code and develop reusable components that can be leveraged across use cases.
• Troubleshoot production issues across model, retrieval, tool-integration, and post-processing layers, driving root-cause analysis and preventative fixes.
• Embed Responsible AI practices, guardrails, and governance controls into delivery and operations to support auditability, traceability, and well-controlled execution.
• Produce recurring health views and reporting for key performance indicators such as adoption, accuracy, and override or exception rates to support stakeholders.
• Facilitate requirements elicitation with operations users through workshops, process walkthroughs, and shadowing, translating needs into requirements, controls, and acceptance criteria.

Required qualifications, skills, and capabilities:

• Demonstrate at least 5 years of hands-on experience in applied AI, machine learning, or AI-powered automation, including delivery of production or production-like solutions.
• Show strong proficiency in Python, including asynchronous programming, to build clean, maintainable, well-tested code.
• Apply Large Language Model techniques, including prompt engineering, structured or JSON-schema outputs, and robustness methods for real-world tasks.
• Use LangGraph or an equivalent graph-based orchestration framework to build multi-step AI workflows.
• Define and validate schemas using Pydantic and perform data processing using pandas.
• Implement Optical Character Recognition and document digitization workflows, including PDF text extraction, OCR tools such as AWS Textract or equivalent, post-OCR cleanup, validation, and exception handling.
• Integrate external tools and services via application programming interfaces, including tool-using assistant patterns with strong validation and error handling.
• Follow core software engineering practices, including automated testing, continuous integration and continuous delivery, secure development, and production readiness (logging, metrics, tracing, runbooks, incident response participation).
• Use Git and a hosted repository platform to manage branching, pull requests, and code review workflows.
• Define objective success metrics and evaluate solution quality with clear problem framing and effective communication across technical and non-technical stakeholders.
• Analyze adoption and performance drivers (such as accuracy and override patterns) using structured root-cause analysis and measurable remediation.

Preferred qualifications, skills, and capabilities:

• Developing experience using AI-powered analytics, workflow automation, or intelligent process tools to drive efficiency gains, reduce manual effort, or improve accuracy in operational processes.
• Hold a Master’s degree with a specialization in AI or machine learning, or a closely related quantitative field.
• Deliver enterprise Large Language Model-powered or agentic applications with accountability for service health, including service level objectives, reliability, and operational excellence.
• Apply retrieval-augmented generation components, including embeddings, vector stores, retrieval quality evaluation, grounding approaches, and hallucination mitigation.
• Interpret evaluation and statistical concepts such as confusion matrix, precision, recall, F1, error analysis, A/B testing, and statistical significance to measure and defend model performance.
• Use classical machine learning frameworks such as PyTorch, TensorFlow, or scikit-learn as needed.
y>Deploy solutions as scalable, observable backend services or application programming interfaces, including familiarity with containerization and cloud-native deployment.y>