Data Scientist - LLM, Agentic AI & Predictive Modeling

About this role

Become a part of our caring community
The Data Scientist 2 – LLM, Agentic AI & Predictive Modeling is a hands-on role focused on designing, deploying, and scaling predictive models, LLM-based, and agentic AI solutions in enterprise environments. The role combines advanced analytics, machine learning, and generative AI with strong MLOps, cloud deployment, and Responsible AI practices to deliver production-ready solutions that drive measurable business and customer impact.


Key Responsibilities
LLM, Agentic AI & Predictive Modeling

• Identify, design, and implement AI use cases leveraging LLMs, Agentic AI, generative AI, predictive modeling, machine learning, deep learning, and advanced analytics.
• Develop, fine-tune, and deploy LLM-based and agent-based systems for enterprise use cases such as conversational AI, workflow automation, reasoning systems, and decision support.
• Design and deploy predictive models including:

• Classification, regression, and ranking models
• Time series forecasting
• Anomaly and fraud detection
• Churn, propensity, and risk models
• Recommender systems and uplift modeling

• Translate predictive model outputs into actionable business signals, integrating them into downstream systems, dashboards, and AI-driven workflows.

Structured & Unstructured Data Modeling

• Engineer, train, and validate machine learning and deep learning models in Python for both structured and unstructured data, including tabular, text, and image data.
• Apply feature engineering, model calibration, interpretability, and performance optimization techniques to predictive models.
• Combine predictive models with LLMs and agentic systems (e.g., predictive scoring feeding agent decisions or RAG pipelines).

NLP, Multimodal & Computer Vision

• Apply NLP techniques such as text mining, semantic search, sentiment analysis, embeddings, and knowledge graph construction.
• Build and deploy computer vision and multimodal models, including image c...