Quant Modeling Lead (Control Room) - Hyderabad

JPMorgan Chase · Hyderabad

Apply on the company website

Quant Modeling (AI/ML) Lead - Hyderabad

• About Control Management
• Control Managers are responsible for having a deep understanding of the business, its underlying processes and the compliance and operational risk and control environment. Subject matter expertise/innovative tools and technology.

Control Management maintains a strong and consistent control environment across the firm. With Control Managers appointed for each Line of Business, Function and Region, there is a comprehensive coverage and joint accountability model with the business executive that promotes early operational risk identification and assessment, effective design and evaluation of controls and sustainable solutions to mitigate operational risk.

• About the Team
• The Quant Modeling team is an integral part of the Controls Room within Control Management, focused on building AI and data science solutions that strengthen control monitoring, analytics, and decision support. The team develops, trains, fine-tunes, and evaluates deep learning and NLP models, including transformer-based and domain-specific architectures, while also building agentic AI systems that support multi-step reasoning, workflow orchestration, tool use, and data integration.

Ideal Candidate

This role is ideal for a hands-on candidate who can build and train NLP/deep learning models, run experiments, fine-tune and evaluate solutions, and design agentic AI systems that apply reasoning, tool use, orchestration, and data integration to practical control management use cases.

Job Responsibilities

• Deep learning model training and experimentation: Design and execute hands-on experiments to train and fine-tune deep learning models, including transformer-based models, embedding models, and domain-specific NLP architectures.
• NLP model and custom architecture development: Develop NLP solutions for text classification, information extraction, semantic search, and retrieval-augmented generation, while experimenting with neural network layers, attention mechanisms, pooling strategies, classification heads, and task-specific modules.
• Model optimization, loss function design, and evaluation: Improve model performance through hyperparameter tuning, custom loss functions, systematic error analysis, and benchmarking using technical and business-aligned metrics such as accuracy, precision, recall, F1, robustness, and latency.
• Agentic AI system design and architecture: Design and implement multi-agent systems that support multi-turn reasoning, tool use, dynamic decision-making, agent decomposition, memory management, and multi-agent coordination for control management problems.
• Agent orchestration, tooling, and data integration: Build orchestration logic for agent routing, state management, workflow coordination, and integration with tools and data sources that agents use to retrieve information, validate data, and take action.
• LLM behavior, evaluation, and monitoring: Shape agent behavior through prompt design and reasoning patterns, and implement evaluation, observability, and monitoring for accuracy, reliability, reasoning traces, tool calls, and decision auditing.
• Reusable agentic AI patterns and governance: Build reusable agent templates, patterns, and documentation that accelerate delivery, while supporting traceability, human-in-the-loop checkpoints, safeguards, and operational procedures.
• Training data strategy for NLP and deep learning: Partner with control managers, domain SMEs, data owners, and technology teams to understand workflows, decision points, and actionable outcomes; source relevant structured and unstructured data; and curate, label, and validate training and evaluation datasets.
• People management aligned to delivery above: Lead and develop a team delivering these systems by setting priorities, coordinating work across model, agent, tool, and orchestration streams, providing technical guidance, and ensuring quality, reliability, documentation, and delivery standards.

Required qualifications, capabilities, and skills

 

• BS/MS/PhD in Computer Science, Data Science, Engineering, or a related quantitative field.

• 15+ years of experience building production or near-production software, data, ML, or backend systems.

• Strong hands-on Python skills, including debugging, testing, and writing maintainable production code.

• Hands-on experience building, training, fine-tuning, evaluating, and deploying AI/ML, NLP, or deep learning models.

• Strong understanding of LLMs, transformers, embeddings, and generative AI concepts.

• Hands-on experience with at least one LLM or agentic AI framework.

• Experience building LLM-based or agentic AI systems with tool use, orchestration, state management, routing, and multi-step reasoning.

• Experience integrating APIs, tools, data services, and enterprise systems into AI/ML or software solutions.

• Experience implementing agent memory, context management, and multi-turn conversation patterns 

• Strong independent problem-solving skills and ability to work through ambiguity.

• Clear communicator and effective cross-functional collaborator.

Preferred qualifications, capabilities, and skills

• Production/enterprise experience building or deploying agentic systems
• Familiarity with production monitoring, logging, error handling, and operational support.

• Familiarity with responsible AI, model governance, and human-in-the-loop review patterns
• Experience with observability and tracing for agent behavior/reasoning
• Experience in risk, controls, compliance, or other highly governed environments

• Location – Hyderabad
• Shift – EMEA

More jobs at JPMorgan Chase · Quant Research jobs in Hyderabad