Capgemini - Generative AI Engineer

GenAI Engineer

Job Description :

GenAI Engineers will build and deliver components of Generative AI and Agentic AI solutions, leveraging Large Language Models (LLMs) and Small Language Models (SLMs) across enterprise use cases. Working within a broader project team and under the guidance of senior engineers and architects, the GenAI Engineer implements well-defined features and technical designs, experiments with new GenAI techniques, and builds strong hands-on depth in fine-tuning and deploying LLM/SLM-based solutions. The role calls for curiosity, a strong learning mindset, and the ability to independently execute well-scoped tasks.

Candidate Requirements :

Education :

- Bachelor's or Master's degree in Computer Science, Data Science, Artificial Intelligence, or a related field.

Experience :

- 5+ years of experience in AI/ML engineering, including at least 1 to 2 years of hands-on experience building GenAI/LLM-based applications and SLM fine tuning to create specialized models.

- Experience contributing to at least one GenAI use case through to production deployment as part of a project team.

- Experience on specialized model training and deployment on sovereign/local on premise infrastructure.

Skills :

Required Skills :

- Hands-on experience fine-tuning and adapting LLMs and SLMs using standard approaches, including LoRA, QLoRA, PEFT, and instruction tuning, under the direction of senior engineers or architects.

- Working knowledge of commonly used LLMs (e.g., GPT-4/4o, Claude, Gemini, Llama, Mistral) and SLMs (e.g., Phi, Gemma, Mistral-7B) for well-defined enterprise use cases.

- Proficiency in Python and GenAI development frameworks such as LangChain, LlamaIndex, Semantic Kernel, Haystack, AutoGen, and CrewAI.

- Practical experience with prompt engineering, prompt chaining, and few-shot/zero-shot techniques.

- Experience implementing components of Agentic AI systems, including tool/function calling and agent workflows, based on architecture and patterns defined by senior team members.

- Experience building Retrieval-Augmented Generation (RAG) pipelines, including chunking, embedding models, and vector databases (e.g., Pinecone, Weaviate, FAISS, Chroma, Milvus).

- A strong learning and innovation mindset, with the ability to experiment with new GenAI techniques and bring ideas to the team.

- Good analytical and problem-solving skills, with the ability to independently execute well-scoped tasks and communicate progress clearly.

Preferred Skills :

- Familiarity with responsible AI practices, including guardrails, hallucination mitigation, and bias detection.

- Exposure to cloud GenAI platforms such as Azure OpenAI, AWS Bedrock, or Google Vertex AI.

- Basic knowledge of MLOps/LLMOps practices, including CI/CD for ML, model versioning, and monitoring tools (e.g., LangSmith, Weights & Biases).

- Understanding of containerization tools (e.g., Docker) for packaging GenAI services.

- Exposure to building or consuming REST/GraphQL APIs for GenAI capabilities.

- Familiarity with knowledge graphs and semantic search.

Responsibilities :

- Develop and fine-tune GenAI/Agentic AI components using LLMs/SLMs against defined technical designs.

- Build and support RAG pipelines, vector search, and retrieval components.

- Implement agent workflows, including tool use and task execution, based on architecture defined by senior engineers.

- Assist in fine-tuning and optimizing LLMs/SLMs under the guidance of senior team members.

- Contribute to the delivery of assigned GenAI features within a broader project team, from build through deployment.

- Collaborate with senior engineers, data engineers, and architects to implement solution requirements.

- Support the evaluation of new GenAI tools and frameworks.

- Apply guardrails and evaluation checks as defined by the team's responsible AI practices.

- Test and validate GenAI solution components for accuracy and performance.

- Stay current with GenAI, LLM, SLM, and Agentic AI developments and share learnings with the team.