Deep Learning Senior Engineer

O projekcie: At Virtusa ( Former ITMAGINATION), every innovator has the potential to transform and lead in a digital world—but unlocking that potential takes more than technology; it takes a trusted partner who combines engineering excellence, creativity, and an AI-first mindset. Together, we co-create solutions that help businesses grow faster, operate smarter, and make experiences better with technology. We are seeking a highly technical Deep Learning Senior Engineer (T3) to join our AI delivery hub in Poland. In this role, you will be a primary architect and builder of advanced Generative AI solutions, moving past basic wrappers to design sophisticated Deep Learning architectures. You will specialize in Transformer-based models, RAG systems, and LLM orchestration, specifically leveraging the Google Gemini ecosystem on Vertex AI. This is a high-impact role requiring a blend of scientific rigor and production-grade engineering to deliver state-of-the-art AI applications for our global enterprise clients. Wymagania: - 6–8 years of experience in Software Engineering or Machine Learning, with a minimum of 3 years focused on Deep Learning and NLP.   - Expert-level Python skills, including deep proficiency with scientific and AI libraries (NumPy, Pandas, PyTorch, or TensorFlow).   - Strong theoretical and practical understanding of Transformers, attention mechanisms, and semantic embeddings.   - Proven track record of building production-ready applications with LangChain, LlamaIndex, and LLM APIs (OpenAI, Anthropic, or Vertex AI).   - Hands-on experience with FAISS, Pinecone, or Weaviate, including indexing strategies, metadata filtering, and hybrid search optimization.   - Advanced experience with GCP, specifically Vertex AI (Model Garden, Pipelines, and Notebooks) and Cloud Storage.   - Deep understanding of NLP concepts such as tokenization, named entity recognition (NER), and semantic search logic.   - Experience using RAGAS or similar tools to quantify model performance (precision, recall, faithfulness).   - Practical knowledge of Docker and Kubernetes; familiarity with CI/CD for ML models and automated deployment workflows.   - Experience building scalable REST/GraphQL APIs and microservices for AI-driven applications.   - Understanding of Responsible AI practices, including data privacy (GDPR), PII masking, and bias detection in LLM outputs.   - Strong analytical problem-solving skills and the ability to work in an Agile (Jira) environment.   - Professional English (C1) is mandatory for collaboration with our international technical leadership.   - Bachelor’s or Master’s degree in Computer Science, AI, Mathematics, or a related quantitative field. Codzienne zadania: - Model Architecture & Design: Design and implement high-performance Generative AI applications utilizing Transformers, Diffusion models, and advanced NLP techniques. - LLM Orchestration: Build and manage complex, agent-based workflows using frameworks like LangChain and LlamaIndex to automate multi-step reasoning tasks. - Advanced RAG Systems: Architect end-to-end Retrieval-Augmented Generation (RAG) pipelines, integrating enterprise data with Vector Databases (Pinecone, FAISS, Weaviate) while ensuring high semantic relevance. - Google GenAI Mastery: Lead the implementation of Google Gemini models within the Vertex AI platform, optimizing for latency, throughput, and cost. - Fine-tuning & Optimization: Perform model fine-tuning, quantization, and embedding optimization to tailor LLMs to specific domain requirements and enterprise datasets. - Prompt Engineering & Evaluation: Design sophisticated prompt strategies and implement rigorous evaluation frameworks (e.g., RAGAS) to track model accuracy, hallucination rates, and drift. - Deployment & Scaling: Collaborate with MLOps teams to deploy models into production environments using Docker and Kubernetes, ensuring scalability and fault tolerance.