Senior Applied AI Scientist

Kai is the AI company rebuilding cybersecurity for the machine-speed era. Founded by second time founders and trusted by Fortune 500 enterprises, Kai is building a future where security has no categories, no silos, and no human speed bottlenecks. The Kai Agentic AI Platform replaces fragmented, human-limited workflows with agentic AI systems that continuously contextualize, assess, reason, and execute security work at machine speed - making human defenders, superhuman.

Why Join Kai

• Well-funded: With $125M raised, we have the capital, runway, and resolve to rebuild cybersecurity from first principles.
• Proven: We've earned the trust of Fortune 500 and Global 1000 companies, and we're just getting started. Their confidence in Kai reflects what we've built: an AI-powered cybersecurity platform that performs at the scale and speed the enterprise demands.
• Experienced founders: Our founding team consists of second-time entrepreneurs, each with over 20 years of experience in the cybersecurity industry. Their proven expertise and vision drive our ambitious goals.
• World-class leadership team: Our Heads of AI, Engineering, and Product bring extensive experience from some of the world’s most influential companies, ensuring top-tier mentorship, direction, and vision.
• Frontier AI Applied Research Team: Our researchers operate at the leading edge of agentic AI systems, translating breakthrough capabilities into real-world cybersecurity applications.
• Generous compensation: We offer highly competitive salaries, equity options, and a supportive work environment. Your contributions will be valued and rewarded as we grow together.
  
  

Key Responsibilities

Core Applied AI Research

• Collaborate with cybersecurity researchers and stakeholders to scope AI-driven solutions to security problems (e.g., vulnerability management, code analysis, threat detection).
• Conduct applied research using the latest LLMs and embedding models (Claude, Google GenAI, Unsloth, vLLM).
• Prototype, fine-tune, and evaluate GenAI and RAG/CAG architectures for classification, summarization, reasoning, and context synthesis.
• Perform embedding-level optimization for text, code, and image data using Unsloth, Hugging Face, Voyage, or similar frameworks.
  
  

System Development & Integration

• Develop and test end-to-end AI pipelines integrating Milvus or Pinecone for semantic retrieval.
• Build agentic AI systems using LangGraph or similar frameworks to enable autonomous reasoning and task chaining.
• Collaborate with MLOps engineers to deploy and monitor AI models in production securely and efficiently.
• Contribute to synthetic data generation pipelines for fine-tuning and evaluation.
  
  

Evaluation & Optimization

• Implement evaluation frameworks using DeepEval and GenAI tools (Claude / Google GenAI) for factuality, reliability, and robustness.
• Optimize model performance across latency, accuracy, and cost using vLLM, quantization, or caching strategies.
• Maintain reproducible experiment tracking with MLflow, Weights & Biases, or internal tools.
  
  

Innovation & Leadership

• Stay ahead of GenAI trends — multi-modal reasoning, agentic orchestration, embedding adaptation.
• Explore hybrid LLM deployment strategies (local Unsloth/vLLM cloud APIs like Claude, Google GenAI).
• Document best practices, share learnings, and mentor junior scientists on applied GenAI workflows.
  
  

Qualifications

Required

• 4 years in Applied AI / Machine Learning Research / Data Science.
• Strong understanding of LLMs, embeddings, RAG systems, and multimodal learning.
• Proficiency in Python and frameworks like PyTorch, Transformers, Hugging Face, or LangChain.
• Experience in prompt engineering, model evaluation, and retrieval-based reasoning.
• Hands-on experience with vector databases (Milvus / Pinecone) and orchestration frameworks (LangGraph / LangChain).
• Strong communication skills and ability to collaborate across research and engineering functions.
  
  

Preferred

• Experience with fine-tuning LLMs or embeddings using Unsloth or similar frameworks.
• Familiarity with Claude / Google GenAI APIs for cloud-based inference and evaluation.
• Exposure to cybersecurity or enterprise data (CVEs, pluginText, network or asset logs).
• Prior work on synthetic data generation and evaluation frameworks (DeepEval).
• Experience in a fast-paced startup or applied research environment.