Position Overview:
We are currently hiring a Full Stack AI Engineer to support work at large, East-Coast based bank within their AI Cybersecurity Team.
Depth & Scope:
· Using Azure AI services to deploy AI models/applications – Azure OpenAI, Azure Machine Learning, Azure AI Studio
· Full Stack Engineering background
o Linux, Windows Sys Admin, Middleware, disciplines, Network Engineering, and some development
o Someone who can truly navigate the full stack
o Understands hardware to application layer
o How to wire AI engines into this environment, how to build the architecture
o Deployments in Azure
· Strong Prompt Engineering, Python Development and REST API skills
· Understanding of different GenAI LLMs, development frameworks (Langchain, Semantic Kernel), and AI ecosystem tooling (evaluation, monitoring, governance, security, vector databases)
· Ability to identify and design approaches for creating guardrails to mitigate GenAI security/cyber risks
· Deep understanding of major components of RAG applications, usage patterns, and optimization
· LLM Fine-tuning approaches
· Understanding the full lifecycle of AI model development, deployment
· Tools for support large scale Inferencing
· Solid understanding of machine learning algorithms, neural network architectures, and LLM model training processes
· Knowledge of different Agentic frameworks
· Familiarity with g Azure AI services – Azure OpenAI, Azure Machine Learning, Azure AI Studio
· Solid understanding of machine learning algorithms, neural network architectures, and AI model training processes
· Experience with adversarial machine learning techniques and familiarity with common attack vectors against AI systems
· Familiarity with AI ethics and bias in AI systems
· Knowledge of techniques for interpreting and explaining AI model decisions
· Ability to identify and design approaches for creating guardrails to mitigate GenAI security/cyber risks & adversarial attacks
· Knowledge of OWASP and NIST GenAI security guidelines & policies
· Understanding the different areas of evaluation for a RAG system
· LLM model evaluation tooling (e.g. Using LLMs-as-a-judge)
