We are building a next-generation Autonomous Security Platform. While our AI Agent Engineers focus on workflow orchestration, YOU will build the Senses and the Brain of the system.
This is a unique hybrid role where you bridge two worlds:
1. The Senses (Detection Core): Building high-precision Anomaly Detection models that run on bothEdge (Endpoints) and Cloud to filter noise & spot threats.
2. The Brain (Agent Intelligence): Fine-tuning specialized LLMs to equip our AI Agents with deepcybersecurity understanding (Log analysis, Threat intelligence reasoning)
Key Reponsibilities
Distributed Anomaly Detection:
- Design lightweight, privacy-first models (e.g., Autoencoders, OCSVM) deployed on customer endpoints via ONNX/TFLite. Your challenge is to detect anomalies in real-time with a negligible CPU/Memory footprint.
- Build heavy-duty Deep Learning models (Graph Neural Networks, Transformers for Time-series) to analyze complex behavioral patterns (UEBA) and lateral movement across the entire network.
- Apply advanced Model Compression techniques (Quantization, Pruning, Knowledge Distillation) to bridge the gap between Research Accuracy and Production Performance.
Fine-Tuning Models for AI Agents:
- Lead the initiative to fine-tune open-source LLMs (Llama 3, Mistral, Qwen) using techniques like LoRA/QLoRA.
- Create specialized Expert Models that AI Agents can call upon to:
+ Understand raw logs (Splunk/JSON) better than generic models.
+ Summarize complex Incident Reports.
+ Reason for attack tactics (MITRE ATT&CK) with reduced hallucination
- Build pipelines to convert unstructured Threat Intelligence and historical incident logs into high-quality instruction datasets for training.
MLOps & Collaboration:
- Expose your Detection Models (e.g., Risk Score) and Fine-tuned LLMs as reliable APIs that the Agent System can consume during automated investigations.
- Own the pipeline for continuous training, evaluation, and drift monitoring to ensure models adapt to ever-changing cyber threats.
Research & Innovation:
- Stay up to date on the latest data science and machine learning research.
- Explore and evaluate new techniques and technologies.
Qualifications
Must have:
- Education: Bachelor's degree/ Master's degree or Ph.D. in Computer Science, Artificial Intelligence, Machine Learning, Electrical Engineering, or a related field.
- 5+ years of experience in Data Science or Applied Machine Learning.
- StrongproficiencywithPyTorchor TensorFlow and the Transformer architecture.
- Experience deploying models into production.
- Solid understanding of Statistics, Linear Algebra, and Optimization algorithms.
- Experience with Anomaly Detection (Unsupervised Learning) and handling highly Imbalanced Data.
- Hands-on experience with LLM Fine-tuning (HuggingFaceecosystem, PEFT/LoRA).
- Familiarity with model optimization tools (ONNX,TensorRT,TFLite).
Nice to have:
- Domain knowledge in Cybersecurity (Log analysis, Malware, Network traffic).
- Experience with Graph Neural Networks (GNN) or Big Data tools (Spark/Kafka).
