The Role
Develop OXMIQ AI Infrastructure Management Software using state of the art agentic development flows.
Objectively evaluate LLMs for incorporation into this architecture.
Develop inference containers and SW stacks for the most popular models on hugging face.
Responsibilities:
- Develop software in python for OXMIQ AI Infrastructure Management System
- Build MCP Tools in python and javascript
- Develop software to interface with k8s instances
- Develop authentication and authorization systems for use with Enterprise auth systems such as Microsoft Entra ID, Google Oauth, etc.
- Evaluate and choose appropriate models for various agentic tasks
- Work on inference software stacks including vLLM, llama.Cpp, sglang, etc.
- Develop software that can process models stored in registries such as huggingface
- Develop Containers that include complete inference stacks
Qualifications:
- 5+ years of python development
- 1 year of work with Large Language Model ecosystems
Required Skills:
- Deep understanding of python
- Agentic loops built using modern LLMs without the use of libraries such as langgraph, smolagents
- Knowledge of LLM inference stacks
- Expert in producing polished production quality software using Code Generation tools
- Expert in using Code Generation tools to generate automated tests
- Knowledge of LLM hosting software such as vLLM, sglang and llama.Cpp
- Knowledge of Linux server systems
- Kubernetes knowledge, especially in the Cloud