About the Role
We are seeking an experienced Machine Learning Lead to helm our Machine Learning team. In this pivotal role, you will be the engineering architect behind Vulcan’s core AI capabilities. You will act as the nexus between Research, Platform, and Product. Your mission is to translate cutting-edge findings on GenAI threats into robust, production-ready machine learning models that power our GenAI Security Guardrails (Blue Team) and Automated Vulnerability Assessment (Red Team).
Crucially, you will serve as the bridge between deep tech and business strategy, articulating technical constraints (like FLOPS and latency) to leadership and clients while guiding the engineering direction.
Key Responsibilities
1. Model Development & Optimization (Training & Fine-tuning):
- Research to Production: Collaborate with the Security Research Team to operationalize new threat detection techniques. They identify the "what" (e.g., new prompt injection patterns); you determine the "how" (model architecture, training strategy).
- Fine-tuning & Adaptation: Lead the fine-tuning of Language Models (e.g., using LoRA/PEFT) to optimize for our supported muti-lingual languages and specific security intents.
- Multimodal Readiness: Prepare the system for Multimodal (Text + Image/Audio) capabilities. Evaluate and implement models to detect visual prompt injections and non-textual threats as the product evolves.
2. MLOps& Data Infrastructure:
- Enhance & Scale MLOps: Take ownership of our existing ML pipelines. Focus on optimizing and scaling CI/CD/CT workflows to improve training efficiency and deployment velocity.
- Data Governance: Implement and enforce rigorous Data Versioning strategies (e.g., DVC) to ensure complete reproducibility of model artifacts and datasets.
- Monitoring & Reliability: Maintain rigorous monitoring for model drift and performance, ensuring high reliability in a production security environment.
3. Cross-Functional Implementation & Leadership:
- Platform Collaboration: Work closely with the Platform Engineering Team to integrate ML models into the broader product architecture. Ensure seamless interaction between model inference services and the main platform logic.
- Team Leadership: Lead and mentor Machine Learning Engineers, fostering a culture of engineering rigor, code quality, and operational excellence.
- Resource Management: Manage GPU resources and compute budgets effectively for both training and inference workloads.
4. Technical Strategy & Stakeholder Management:
- Translating Tech to Business: Act as the technical voice of the ML team. You must effectively explain complex ML concepts (e.g., FLOPS, quantization trade-offs, model latency vs. accuracy) to executive leadership and clients.
- Cost-Benefit Analysis: Justify compute resource investments. Articulate the trade-off between infrastructure costs (GPU hours) and performance gains to non-technical stakeholders.
Qualifications
- Experience: 5+ years in Machine Learning Engineering, with specific experience in leading technical projects or mentoring engineers.
- Communication & Business Acumen: Exceptional ability to distill complex technical topics (e.g., compute complexity, infrastructure costs) into clear, business-relevant insights for decision-makers.
- MLOps Proficiency: Proven experience in optimizing ML pipelines and infrastructure. Familiarity with tools like MLflow, Kubeflow, Airflow, and Data Versioning tools (DVC, etc.).
- Engineering First: Proficient in Python, Docker, and Kubernetes. You treat ML models as software artifacts that need testing and version control.
- NLP & LLM Expertise: Experience with Transformer architectures, Embeddings, and LLM fine-tuning. Familiarity with frameworks like PyTorch, Hugging Face, and vLLM.
- Language Support: Experience processing or fine-tuning models for multi-lingual environments.
Nice to Have
- Multimodal Expertise: Experience working with Multimodal models (Image-to-Text, Text-to-Image, VLMs like CLIP, LLaVA).
- Security Awareness: Understanding of GenAI security threats (e.g., Prompt Injection).
- High-Performance Computing: Experience optimizing inference speed (quantization, distillation, vLLM) for real-time applications.
- Vector Database: Experience with Vector DBs for RAG applications.