DCode Talent
Job Summary
Contribute to building a production-ready Generative AI platform for Johnson & Johnson Innovative Medicine as part of an emerging R&D team. Design and implement scalable Retrieval-Augmented Generation (RAG) pipelines for large-scale document processing and semantic chunking from cloud sources (Azure Blob, AWS S3). Integrate and tune embedding models and vector databases (e.g., Milvus) for high-performance document retrieval. Develop hybrid retrieval systems (BM25 and vector search) and optimize retrieval performance using key metrics (MRR, recall@k). Apply and fine-tune large language models (LLMs) for complex NLP tasks such as named-entity recognition, question answering, and summarization. Build and optimize prompt engineering solutions, ensuring structured outputs and reducing model hallucinations. Develop, containerize, and deploy agent-based microservices (FastAPI, Azure Functions); define Infrastructure as Code (Terraform/ARM). Establish CI/CD workflows (GitHub Actions) for automated testing and deployment, including monitoring and alerting for system performance and SLA compliance. Profile and optimize system performance and cost (batching, caching, early-stopping), while producing high-quality documentation and architecture diagrams. Ensure security and compliance with industry regulations (HIPAA/GxP), including data encryption and PII redaction. Collaborate onsite three days a week (hybrid role; must be within driving distance). No visa sponsorship available for this position.
#J-18808-Ljbffr
Contribute to building a production-ready Generative AI platform for Johnson & Johnson Innovative Medicine as part of an emerging R&D team. Design and implement scalable Retrieval-Augmented Generation (RAG) pipelines for large-scale document processing and semantic chunking from cloud sources (Azure Blob, AWS S3). Integrate and tune embedding models and vector databases (e.g., Milvus) for high-performance document retrieval. Develop hybrid retrieval systems (BM25 and vector search) and optimize retrieval performance using key metrics (MRR, recall@k). Apply and fine-tune large language models (LLMs) for complex NLP tasks such as named-entity recognition, question answering, and summarization. Build and optimize prompt engineering solutions, ensuring structured outputs and reducing model hallucinations. Develop, containerize, and deploy agent-based microservices (FastAPI, Azure Functions); define Infrastructure as Code (Terraform/ARM). Establish CI/CD workflows (GitHub Actions) for automated testing and deployment, including monitoring and alerting for system performance and SLA compliance. Profile and optimize system performance and cost (batching, caching, early-stopping), while producing high-quality documentation and architecture diagrams. Ensure security and compliance with industry regulations (HIPAA/GxP), including data encryption and PII redaction. Collaborate onsite three days a week (hybrid role; must be within driving distance). No visa sponsorship available for this position.
#J-18808-Ljbffr