Senior AI Engineer
Python
Senior
Agentic Ai Engineer
Ai Ml
Artificial Intelligence
Big Data
Bigdata
Cloud Data Engineering
Cloud Data Platform
Cloud Data Warehouse
Cloud Data Warehouse
Cloud Platform
Cloud Platforms
Data Analysis
Data Engineer
Data Engineering
Data Lakehouse
Data Pipeline
Data Platform
Data Processing
Data Warehouse
Databricks
Generative AI
Generative Ai Engineering
Lang Graph
Large Language Models
Machine Learning
Programming
Programming Language
Programming Languages
Rag Architectures
Software Development
Job Description
Design and operationalize agentic AI for Firm Risk Management in a governed enterprise environment.
Responsibilities
- Design and build modular, maintainable Python applications, APIs, and services integrating LLMs, enterprise data, and business tools.
- Develop LangGraph agent workflows with explicit state, conditional routing, tool execution, checkpointing, and human-in-the-loop approval or escalation when needed.
- Engineer agent context by selecting and assembling relevant instructions, retrieved information, tool outputs, and conversation history within token and access-control constraints.
- Design, version, and test prompts with structured outputs, grounding instructions, and safe handling of incomplete or conflicting information.
- Implement short-term and long-term agent memory strategies including persistence, retrieval, summarization, retention, and isolation between users and sessions.
- Build evaluation frameworks using representative datasets and repeatable regression tests with automated and human evaluation.
- Assess answer quality across grounding, tool-call correctness, task completion, and failure behavior.
- Implement LLM and agent observability using Galileo or comparable platforms (execution traces, quality signals, latency, token usage, cost, and failure analysis).
- Apply guardrails for prompt injection, unsafe tool use, sensitive-data exposure, and unsupported outputs with access controls and auditable execution paths.
- Collaborate with risk specialists, product owners, data engineers, and platform teams to translate requirements into tested applications and support deployment and troubleshooting.
- Participate in technical and code reviews, documentation, and engineering guidance while staying active in implementation.
Requirements
- Strong hands-on Python engineering: application design, asynchronous programming, APIs, automated testing, error handling, and maintainable code.
- Demonstrated experience implementing AI agents with LangGraph, including stateful orchestration and integration with external tools or services beyond chatbots, tutorials, or conceptual designs.
- Practical understanding of LLMs, embeddings, retrieval-augmented generation (RAG), vector search, and retrieval-context tradeoffs.
- Hands-on context engineering, prompt engineering, and agent memory management.
- Experience designing and implementing evaluation frameworks for LLM/agent applications, including test-set construction and interpreting evaluation results.
- Experience with Galileo or a comparable LLM/agent evaluation and observability platform, using traces to debug failures and improve applications.
- Experience delivering and supporting deployed applications, including security, reliability, version control, CI/CD, and production troubleshooting.
- Ability to explain technical decisions, distinguish personal contributions from team outcomes, and collaborate with business and engineering stakeholders.
Technologies
- Python
- LangGraph
- LLMs
- Embeddings
- Retrieval-Augmented Generation (RAG)
- Vector Search
- Galileo
- Snowflake
- Databricks
- Model Context Protocol (MCP)
- GitHub Copilot
- Claude Code
- Amp
- CI/CD
- Containers
Highly Desirable
- Knowledge of Model Risk Management, model development/validation lifecycles, model documentation, governance, or model monitoring.
- Experience in broader risk management, especially credit risk, market risk, or financial-services risk workflows.
- Experience with Snowflake, Databricks, lakehouse architectures, enterprise data pipelines, or governed knowledge integration.
- Familiarity with Model Context Protocol (MCP), cloud deployment, containers, and enterprise identity/access controls.
- Prior experience using AI coding assistants such as GitHub Copilot, Claude Code, or Amp with disciplined review, testing, and secure handling of code and data.
Location: New York, NY (onsite)
Compensation: USD 90 - 90 per hourly