Job Description
Join the Future of Intelligence at Nexus Horizon Systems.
We are on a mission to define the roadmap for Artificial General Intelligence (AGI) by 2026. As a Senior Generative AI Engineer, you will architect and deploy state-of-the-art Large Language Models (LLMs) that power the next generation of enterprise solutions. You will work in a high-performance environment where innovation meets scalability.
Why Join Us?
• Work on cutting-edge LLM fine-tuning and RAG pipelines.
• Competitive equity package and top-tier healthcare.
• Flexible remote-first culture with HQ in the heart of SF.
Responsibilities
- Model Architecture: Design and implement scalable neural network architectures for Large Language Models (LLMs) and multimodal systems.
- Optimization: Apply quantization, pruning, and distillation techniques to optimize model inference speed and reduce latency.
- RAG Systems: Build robust Retrieval-Augmented Generation pipelines to ensure factual accuracy and context-awareness.
- Research & Development: Stay at the forefront of AI research, experiment with novel transformer architectures, and publish findings.
- Collaboration: Partner with Product Managers and Data Scientists to translate complex AI capabilities into user-friendly features.
- Mentorship: Guide a team of junior engineers and researchers, fostering a culture of technical excellence.
Qualifications
- Experience: 5+ years of professional experience in AI/ML engineering with a focus on deep learning frameworks.
- Technical Skills: Proficiency in Python, PyTorch, or TensorFlow. Experience with Hugging Face Transformers is required.
- LLM Expertise: Deep understanding of GPT architectures, BERT variants, and fine-tuning methodologies (LoRA, QLoRA).
- MLOps: Familiarity with MLOps tools (Docker, Kubernetes, MLflow) and cloud platforms (AWS/GCP/Azure).
- Education: MS or PhD in Computer Science, Machine Learning, or a related quantitative field.
- Problem Solving: Exceptional ability to debug complex distributed systems and optimize training pipelines.