New opportunity

Data Scientist - Associate

JPMorgan Chase · Mumbai, India

About this role

You are a strategic thinker passionate about driving solutions in Applied AI & Machine Learning. As an Applied AI & Machine Learning Professional within our Asset Management Data Science Team, you will be responsible for setting and improving our organizational objectives and ensuring their consistent accomplishment.

Responsibilities

Design and implement scalable architecture for LLM-powered investment tools, including reusable patterns for integration, governance, observability, and access control.
Build and scale AI applications and automated workflows using models, retrieval, and tools with orchestration for state management, retries, checkpointing, deterministic controls, and appropriate human oversight.
Develop context-management and retrieval-augmented generation (RAG) capabilities, including ingestion, chunking, metadata, embeddings, hybrid search, reranking, and grounded outputs.
Create reliable data and knowledge pipelines that transform structured and unstructured enterprise content into governed, high-quality inputs for AI applications.
Establish engineering standards for reusable tools and model integrations, including interfaces, permissions, testing, failure handling, and documentation.
Implement evaluation and end-to-end observability for AI systems, including structured logging and tracing across model calls, retrieval steps, and tool execution.
Optimize AI systems for quality, task completion, latency, token usage, cost, reliability, and business impact.
Partner with portfolio managers and research teams to understand investment processes and translate them into practical solutions within the firm’s ecosystem.

Required qualifications

Earn a Master’s degree or PhD in computer science, statistics, mathematics, engineering, econometrics, or another quantitative field.
Apply a strong foundation in statistics, probability, experimental design, and machine learning to select appropriate methods and interpret quantitative results.
Build, deploy, and scale LLM-powered applications, including retrieval, tool use, workflow orchestration, state management, structured outputs, and evaluation, using frameworks such as LangGraph, Semantic Kernel, LlamaIndex, or equivalent.
Design prompts and context effectively, including embeddings, vector and keyword search, reranking, model selection, and quality/latency/cost optimization.
Write strong Python and SQL, using common data and machine learning libraries and frameworks.
Demonstrate sound numerical intuition and understanding of financial markets, investment research, portfolio construction, risk, performance, and investment data.
Translate ambiguous business requirements into scalable technical solutions and communicate technical trade-offs to a range of stakeholders.

Preferred qualifications

Deliver generative AI or machine learning solutions in a regulated enterprise environment.
Bring front-office or buy-side experience, particularly in investment research, portfolio analytics, performance attribution, or decision analytics.
Incorporate unstructured or alternative data into research and production workflows.
Apply familiarity with Model Context Protocol (MCP) or comparable standards for securely connecting AI applications to enterprise data, tools, and services.
Use LLM observability and evaluation standards or platforms such as OpenTelemetry, OpenInference, Arize Phoenix, LangSmith, or equivalent (including traces across model, retrieval, and tool-execution steps).
Leverage familiarity with knowledge graphs, multimodal models, fine-tuning, synthetic data, or advanced model-evaluation techniques.
Pursue CFA qualification or progress toward the CFA designation.

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

Skills for this role

PythonSQLMachine LearningLLMRAGLangGraphSemantic KernelLlamaIndexOpenTelemetryOpenInferenceArize PhoenixLangSmithModel Context Protocol

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

Scalable LLM Architecture and Integration
Designing and implementing robust architecture for LLM-powered investment tools, focusing on reusable integration patterns, governance, observability, and strict access control within an enterprise ecosystem.
Advanced Retrieval-Augmented Generation (RAG) and Workflows
Building and scaling AI applications with sophisticated context-management, hybrid search, ingestion, chunking, reranking, and workflow orchestration featuring state management, retries, and checkpointing.
Financial Domain Translation
Partnering directly with portfolio managers and research teams to understand complex investment processes, translating ambiguous business requirements into practical, quantitative technical solutions.
Rigorous System Evaluation and Observability
Implementing comprehensive end-to-end observability, structured logging, and tracing across model calls, retrieval steps, and tool execution to optimize quality, latency, token usage, and cost.

Likely challenges to think through

Balancing Deterministic Controls with LLM Stochasticity
A key technical difficulty is embedding deterministic checks, state management, and human oversight into autonomous agentic or tool-use workflows without introducing excessive latency or degrading model flexibility. Prepare by reviewing orchestration frameworks like LangGraph and implementing checkpointing and retry logic.
Translating Abstract Financial Workflows into Technical Specifications
Bridging the gap between high-level investment research or portfolio management needs and precise technical inputs requires deep numerical intuition. Prepare by studying investment processes, asset management terminology, and practicing how to communicate technical trade-offs clearly to non-technical stakeholders.
Maintaining Strict Governance and Observability in Regulated Environments
Deploying generative AI solutions in a regulated asset management setting demands rigorous traceability, secure data pipelines, and compliance standards. Prepare by exploring enterprise observability platforms like OpenTelemetry, Arize Phoenix, or LangSmith, and secure protocol patterns like Model Context Protocol (MCP).

Practice your answers

Technical practice

Scalable Architecture for LLM-Powered Investment Tools

What to coverDiscuss how you would design a scalable architecture for LLM-powered tools within an asset management environment, focusing on integration patterns, governance, observability, and role-based access control. Address how you ensure low latency while maintaining strict security boundaries around enterprise data.

State Management and Orchestration in AI Workflows

What to coverExplain the role of workflow orchestration frameworks like LangGraph, Semantic Kernel, or LlamaIndex in managing state, retries, checkpointing, and deterministic controls for automated AI agents. Emphasize how human-in-the-loop oversight is integrated into complex multi-step pipelines.

Advanced RAG and Context Management Pipelines

What to coverDescribe the end-to-end design of a retrieval-augmented generation pipeline, covering ingestion strategies, chunking methodologies, metadata tagging, hybrid search (vector plus keyword), and reranking. Explain how grounded outputs are verified to prevent hallucination in financial research workflows.

End-to-End Observability and Evaluation for LLM Systems

What to coverExplain how you implement end-to-end observability and structured tracing across model calls, retrieval steps, and tool execution using platforms like OpenTelemetry, Arize Phoenix, or LangSmith. Discuss key metrics for monitoring quality, task completion, token usage, cost, and reliability in production.

Model Context Protocol (MCP) in Enterprise Integration

What to coverDiscuss the application of the Model Context Protocol (MCP) or comparable standards for securely connecting AI applications to enterprise data sources, external tools, and internal services. Highlight security implications, protocol overhead, and authorization mechanisms.

Financial Market Fundamentals and Quantitative Modeling

What to coverExplain how quantitative foundations in statistics, probability, and experimental design are applied to investment research, portfolio construction, risk management, and performance attribution. Discuss the challenges of incorporating unstructured or alternative data into traditional financial workflows.

Python and SQL Engineering Standards for ML Pipelines

What to coverDescribe your approach to writing robust Python and SQL code for transforming structured and unstructured enterprise content into governed, high-quality inputs for machine learning models. Address code maintainability, unit testing, error handling, and failure management.

Translating Ambiguous Business Requirements into Technical Solutions

What to coverDiscuss a methodology for translating ambiguous requirements from portfolio managers and research teams into scalable technical solutions. Address how you communicate technical trade-offs regarding model latency, cost, and accuracy to diverse stakeholders.

Behavioural practice

Cross-Functional Collaboration with Investment Teams

What to coverDescribe a situation where you partnered with non-technical stakeholders, such as portfolio managers or research analysts, to understand their workflow and deliver an AI or machine learning solution. Explain how you gathered requirements, managed expectations, and ensured the final tool solved their business problem using the situation, action, and result framework.

Balancing Performance, Cost, and Latency in Production

What to coverRecount a project where you had to optimize an AI or machine learning system for competing constraints like latency, token cost, and output quality. Detail the diagnostic steps you took, the trade-offs you evaluated, and the resulting business impact using the situation, action, and result framework.

Navigating Regulatory and Governance Constraints

What to coverShare an experience where you delivered a machine learning or generative AI solution within a regulated enterprise environment with strict data governance and security rules. Describe how you addressed compliance, auditability, and access control requirements using the situation, action, and result framework.

Handling System Failures and Debugging Complex AI Workflows

What to coverDescribe a time when a multi-step AI application or RAG pipeline failed unexpectedly in a production or staging environment. Explain how you utilized tracing and logging tools to isolate the root cause, implemented error handling or retries, and prevented future regressions using the situation, action, and result framework.

Sharpen your knowledge

Topics to prepare

Advanced RAG and Context Management
Revise concepts of ingestion, chunking, metadata enrichment, hybrid search, and reranking to build grounded outputs for enterprise AI applications. Relevant to developing robust retrieval-augmented generation capabilities and handling unstructured investment content.
Financial Markets and Investment Research
Review fundamental principles of portfolio construction, asset management, risk, performance attribution, and investment data workflows. Relevant to partnering effectively with portfolio managers and translating complex investment processes into technical solutions.
AI System Evaluation and Observability
Study end-to-end tracing methodologies, structured logging across model calls, retrieval steps, and tool execution. Relevant to implementing robust evaluation frameworks and optimizing AI systems for quality, cost, and reliability.

Skills to revise

Workflow Orchestration with LangGraph
Practice building, scaling, and managing state for LLM-powered applications using LangGraph, LlamaIndex, or Semantic Kernel, including retries and checkpointing. Relevant to designing automated workflows and deterministic controls requested in the role.

Your preparation checklist

Review Quantitative Foundations and Financial Domain Concepts
Brush up on core statistical concepts, experimental design, and probability, alongside foundational knowledge of financial markets, investment research, portfolio construction, and performance attribution to align with the role's quantitative focus.
Build a Workflow-Orchestrated LLM Application Project
Develop a sample project using frameworks like LangGraph, Semantic Kernel, or LlamaIndex that demonstrates state management, checkpointing, retries, and tool use with human oversight.
Implement an Advanced RAG Pipeline with Observability
Create a retrieval-augmented generation pipeline featuring document ingestion, custom chunking, metadata filtering, hybrid search, reranking, and integration with observability platforms like LangSmith or Arize Phoenix.
Prepare Python and SQL Code Artifacts
Organize code samples demonstrating strong Python and SQL data transformation skills, specifically handling structured and unstructured enterprise content for machine learning pipelines.
Study Enterprise Governance and Secure Protocol Standards
Review concepts related to Model Context Protocol (MCP), secure data connectors, governance, access control, and structured logging in regulated enterprise environments.
Formulate Targeted Questions for the Employer
Draft thoughtful questions regarding how the Asset Management Data Science team partners with portfolio managers, measures AI system latency and cost trade-offs, and handles production observability.

Availability has not yet been verified. 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…