Position Overview Join Virtue AI as a Machine Learning Engineer specializing in Security Applications, where you will contribute significantly to the development and deployment of innovative large models and algorithms aimed at enhancing code security. Your expertise will play a key role in advancing our services and driving industry innovation.
Responsibilities - Develop LLM-based agents focused on various code-security tasks, including red-teaming tests for malware and generating cyber attack simulations, as well as blue-teaming for vulnerability detection.
- Fine-tune custom code-oriented LLMs to optimize their performance while prioritizing reductions in model size.
- Implement efficient inference methods to decrease model latency and improve responsiveness.
- Evaluate and verify the effectiveness of guardrail models and agents.
Requirements We encourage applicants who meet any combination of the requirements below to apply:
- Strong programming skills in languages such as Python and familiarity with LLM libraries including PyTorch, DeepSpeed, OpenRLHF, and vLLM.
- Proven experience in fine-tuning LLMs, specifically customized for code security, such as Llama-coder and Qwen-coder.
- Solid understanding of software security principles, including static and dynamic program analysis, awareness of prevalent security vulnerabilities, and familiarity with common programming languages like Python, C, C , and Java.
- Experience creating LLM-based agents designed for security tasks, encompassing vulnerability detection and system-level defenses like sandboxing and sanitization.