Controls Room AI/ML Developer - Senior Associate
JPMorgan ChaseThe Controls Room is a physical and virtual 'room' containing firmwide control-related information to enable rapid access to relevant data, advanced analytics, and proactive issue identification. The Controls Room Team is designed to facilitate better monitoring and management of the Firm's control environment through the creation and maintenance of Firmwide Systems which support key Operational Risk Frameworks as well as a Firmwide Reporting Utility that provides standardized control-related data and encourages quick, efficient and accurate reporting and analytics. The primary goal is to enhance risk and control oversight through the standardization and automation of operational risk recording and reporting as well as provide access to firmwide aggregated information and to produce business risk insight.
Description for Internal Candidates
Central Control Management solidifies an effective firmwide control framework within and across all lines of business by identifying and remediating control issues with a sense of urgency, irrespective of the functional area. The Central Control Management team works collaboratively with other control disciplines and oversees existing control functions as well as the development of new control functions and protocols. This process enables the firm to engage the appropriate teams in a timely manner and provides the ability to quickly remediate critical control issues across all the impacted areas of the firm.
Help build agentic systems that turn control data into actionable insight through multi-step reasoning, dynamic tool use, and intelligent workflow orchestration. This role offers the opportunity to design and implement intelligent agents that understand control problems, retrieve relevant information, reason across complex scenarios, and take guided action. You will work across data, engineering, and business teams to develop agents that interact naturally with users and systems while maintaining appropriate governance and control. Your work will contribute directly to faster decision-making, smarter automation, and improved risk oversight.
Job summary
As a Controls Room Quant Modeling Senior Associate in the Controls Room team, you design and build agentic systems that improve control monitoring, analytics, and decision support. You will work on agent architecture, workflow orchestration, tool integration, and agent reasoning patterns that enable natural, multi-step problem solving. We are looking for someone with strong engineering fundamentals, practical experience with LLMs and agent frameworks, and the ability to contribute effectively across data, engineering, and business teams.
Job responsibilities
• Design and implement Multi-Agent Dynamic System hat enable multi-turn reasoning, tool use, and dynamic decision-making for control management problems
• Develop and Design and integrate tools and data connections that agents use to retrieve information, validate data, and take action
• Build and maintain orchestration logic for agent routing, state management, and workflow coordination
• Work with LLMs and prompt design to shape agent behavior, reasoning patterns, and response quality
• Contribute to agent architecture and design decisions, including agent decomposition, memory management, and multi-agent coordination
• Partner with data, product, and engineering teams to translate business problems into practical agentic solutions
• Implement validation, testing, and evaluation frameworks for agent performance, accuracy, and reliability
• Support observability and monitoring for agent systems, including reasoning traces, tool calls, and decision auditing
• Help build reusable agent templates, patterns, and tooling that accelerate delivery across use cases
• Document agent design, assumptions, tool specifications, and operational procedures for technical and non-technical audiences
• Support governance and control requirements by implementing traceability, human-in-the-loop checkpoints, and appropriate safeguards into agent workflows
Required qualifications, capabilities, and skills
• Bachelor's degree, Master's degree, or Ph.D. in Computer Science, Data Science, Engineering, or a related quantitative discipline
• 3+ years of experience building software, data, or backend systems in production or near-production environments
• Strong Python programming skills, including debugging, testing, and writing maintainable code
• Foundational understanding of Large Language Models, transformer-based architectures, and LLM capabilities and limitations
• Experience with at least one LLM-based agentic framework or library
• Experience integrating external tools, APIs, and data services into software or machine learning systems
• Experience designing or implementing multi-step workflows, state machines, or orchestration logic
• Experience with prompt engineering, prompt templates, and shaping model behavior through instructions
• Familiarity with monitoring, logging, and operational support practices for production applications
• Ability to solve problems independently within defined scope and collaborate effectively across cross-functional teams
• Ability to communicate technical work clearly to both technical and non-technical stakeholders
Preferred qualifications, capabilities, and skills
• Experience building or deploying agentic systems in production or enterprise environments
• Experience with retrieval-augmented generation (RAG) and grounding strategies for LLMs
• Experience with async task execution, event-driven workflows, or asynchronous coordination patterns
• Experience implementing agent memory, context management, or multi-turn conversation patterns
• Familiarity with responsible AI practices, model governance, or human-in-the-loop review patterns in agent systems
• Experience with observability and tracing tools for understanding system behavior and reasoning
• Experience in risk, controls, compliance, or other highly governed environments
Please note- We are unable to provide sponsorship for this role now or in the future.