PSSM Generative AI Practitioner
Pfizer
- Chennai, Tamil Nadu
- Permanent
- Full-time
- Work with partner development lines to understand the nuances and complexities of scientific fields such as chemistry, physics, or engineering, and identify opportunities for applying generative AI techniques to solve business problems.
- Curate scientific information, experimental record, or other relevant PSSM documents and sources to evaluate LLM performance for pharmaceutical sciences domain-specific knowledge generation.
- Work with system/data engineers and partner lines to build knowledge extraction pipelines across PharmSci development and manufacturing data modalities (e.g., from text documents, images, video, and multimodal datasets).
- Explore transformer models (e.g., GPT, CLIP, ViT-YOLOs) for various tasks, including object content generation, computer vision, and image-text alignment. Evaluate model performance, interpret results, experiment with various training strategies and domain-specific fine-tuning to improve accuracy and efficiency.
- Apply advanced insight retrieval techniques (e.g. RAG, Langchain, Knowledge Graph Embeddings) with generative models for improved information retrieval from PSSM unstructured data and integration of structured data, ontologies, knowledge graphs, or other forms of structured knowledge representation.
- Share findings with stakeholders to improve business decisions and/or influence strategic direction.
- Partner with Pfizer Digital to deploy and maintain AI models in scalable production environments. Monitor model performance, retrain, and optimize models.
- Stay abreast with the latest technologies and industry trends to identify opportunities for enhanced performance and innovation.
- Support upskilling the organization on the application of GenAI and to help build a continuous learning culture.
- Master's Degree or PhD (preferred) in Computer Science, Artificial Intelligence, or a related field.
- Ability to work effectively with large-scale, structured, and unstructured datasets.
- Proficient in prompt design, vector representations and vector databases.
- Solid understanding of transformer architectures, their variations, and generative AI techniques (GANs, VAEs, YOLOS, GPT, etc.).
- Skill in applying domain-specific LLM training and fine-tuning methods.
- Experience with RAG, Langchain, Knowledge Graph Embeddings, and other insight retrieval techniques.
- Proficiency in Python and ML/DL frameworks (PyTorch, TensorFlow)
- Experience in MLOps/LLMops concepts (monitoring, retraining, versioning).
- Strong problem-solving, analytical, research and critical thinking skills.
- Excellent communication, collaboration, and presentation skills.
- Experience of AI applications in pharmaceutical sciences, chemical reactions and chemical representations (preferred but not mandatory).
- Experience with cloud platforms like AWS, Azure, or Google Cloud (preferred but not mandatory).
- Publications at major AI conferences or peer-reviewed journals (preferred but not mandatory).
- Contributions to open-source NLP or computer vision projects (preferred but not mandatory).