Industrial AI Scientist

About the Role

As a Industrial AI Scientist at Avathon, you will play a key role in designing and delivering advanced AI solutions with a strong emphasis on Generative AI, Large Language Models (LLMs) and ML. You will apply scientific rigor to develop scalable, production-ready machine learning systems that drive measurable business impact, working on challenging problems in forecasting, demand planning, renewable energy optimization, Normal behaviour modelling, anomaly detection, and predictive maintenance. You are expected to bring strong expertise in statistical modelling, Advanced ML engineering, and modern AI architectures, particularly in GenAI and LLM-based applications. This role offers the opportunity to work on high-impact projects that shape next-generation AI capabilities within the organization.

The work is equal parts applied research and production delivery. You will develop novel approaches where standard methods fall short and ship them into a platform used by enterprise customers. You will report into the Data Science org and work closely with product and engineering.

You Will

• Design, develop, and deploy machine learning and Generative AI solutions to solve complex business problems
• Build, fine-tune, and optimize Large Language Models (LLMs) and transformer-based architectures for real-world applications
• Apply rigorous scientific methodologies to experimentation, model evaluation, and performance optimization
• Develop scalable ML pipelines and production-grade systems in collaboration with Engineering teams
• Conduct prompt engineering, model alignment, evaluation, and performance benchmarking for GenAI applications
• Work closely with Product, Engineering, and Business stakeholders to translate ambiguous requirements into data-driven AI solutions
• Instrument and monitor LLM applications in production using observability tools, tracking cost, latency, quality, and drift
• Contribute to model governance, responsible AI practices, and performance monitoring in production environments
• Stay current with advancements in Generative AI, LLM research, and applied machine learning, incorporating relevant innovations into company solutions
• Define evaluation metrics, build backtesting pipelines, and run controlled experiments to measure and improve model performance
• Integrate cross-domain signals from Avathon's Computational Knowledge Graph 
• Own the path from research to production in collaboration with ML Engineering, delivering scalable and monitored services
• Mentor junior data scientists on technical depth and AI/ML fundamentals

You'll Have

• Master’s or PhD in Computer Science, Data Science, Statistics, Mathematics, Engineering, or a related quantitative field
• Prior experience applying AI solutions in industrial asset-intensive environments
• Hands-on industry experience in AI engineering, Advance Machine learning, data science, or applied AI roles
• Experience with Generative AI frameworks and Large Language Models (e.g., transformer architectures, fine-tuning, RAG systems)
• Proficiency in Python and modern ML/AI libraries such as PyTorch, TensorFlow, Hugging Face, or equivalent ecosystems
• Understanding of statistical modelling, experimentation, and model evaluation methodologies
• Experience building and deploying Advance ML models into production environments
• Strong problem-solving skills with the ability to work independently on complex and ambiguous problem statements
• Excellent communication skills with the ability to present technical insights clearly to cross-functional stakeholders
• Strong foundation in statistical modeling with sound judgment on when simple methods outperform complex ones
• Experience deploying models into production environments

Preferred Qualifications

• Ph.D. in Operations Research, Industrial Engineering, Statistics, or a related field
• Experience with cloud ML infrastructure (AWS SageMaker, GCP Vertex, or equivalent)
• Domain experience in energy, mining, manufacturing, aerospace, or logistics or Track record of publishing or presenting applied OR/ML work as plus