2 days ago
Bengaluru, IndiaSenior
Responsibilities
- Design and implement single-agent and simple multi-agent workflows for traded-risk teams.
- Lead end-to-end development of agentic AI applications from ideation and data exploration through prototyping and initial deployment.
- Build tool integrations and MCP servers connecting LLM agents to internal risk systems.
- Implement RAG pipelines over risk policies, regulatory guidance, and historical incident logs.
- Develop prompt templates, system instructions, few-shot examples, and a versioned prompt library.
- Build evaluation harnesses measuring accuracy, hallucination rates, and task completion through automated metrics and human review.
- Implement input/output validation, sensitive-data filtering, approval checkpoints, and human-in-the-loop escalation paths.
- Monitor deployed agents for token cost, latency, failure modes, and output-quality drift.
- Document workflows, decision logic, and control points for audit and model-risk review.
- Collaborate with market-risk subject-matter experts to convert manual processes into structured agent tasks.
Requirements
- Working proficiency in Python, including API integration, asynchronous programming, and data manipulation with pandas and NumPy.
- Hands-on experience building agentic applications through professional work, internships, hackathons, or personal projects.
- Practical exposure to an agent orchestration framework such as LangGraph, LangChain, CrewAI, AutoGen, or Semantic Kernel, or direct experience with Claude, OpenAI, or Gemini APIs and function calling.
- Understanding of prompt engineering, including system prompts, structured reasoning prompts, few-shot examples, and JSON or XML output formatting.
- Familiarity with RAG architectures, embeddings, vector databases, chunking strategies, and retrieval evaluation.
- Basic understanding of LLM evaluation techniques such as golden datasets, LLM-as-judge patterns, and regression testing.
- Foundational understanding of market-risk concepts including VaR, Greeks, stress testing, and limit frameworks.
- FRM Part I/II, CQF, or an equivalent certification completed or in progress is relevant.
- Experience developing MCP servers for tool integration is relevant.
- Exposure to model-risk management or SR 11-7-style validation frameworks is relevant.
- Prior experience in a regulated environment such as banking, insurance, or asset management is relevant.
- Background in software engineering, data engineering, or quantitative development is relevant.
- Intern candidates may be considered with one year of AI exposure.
Benefits
- Fixed-term contract role.
- HSBC offers professional development, flexible working, and opportunities for growth.
- Inclusive and diverse workplace with equal-opportunity employment practices.
Tech Stack
Categories
About HSBC
HSBC is a global bank that provides retail, commercial and investment banking, payments, and wealth management to individuals, SMEs, and multinationals. It earns fees, interest income, and trading revenues through branches and digital platforms across major markets. Headquartered in London, it serves over 40 million customers in 58 countries and is listed in London, Hong Kong, New York, and Bermuda.