TCS accelerates AI adoption for enterprise clients (industry, energy, services). We are looking for a passionate Generative AI intern to join our Toulouse team. You will work in a structured project environment (Agile), mentored by an AI Tech Lead, and contribute to GenAI POCs and MVPs that move toward production.

Missions:

• Design and develop AI/GenAI POCs (RAG, agents, assistants, domain copilots) and evolve them into production-ready MVPs.
• Build LLM chains (prompting, tooling, evaluation) and agents (tools, planning, multi-step execution).
• Implement data pipelines for preparation, indexing, and search (hybrid BM25 + vectors).
• Set up evaluation and observability (quality, cost, latency, hallucination rate, tracing).
• Deploy on GCP (preferred) – Vertex AI, Cloud Run, BigQuery, IAM – or, depending on projects, AWS/Azure.
• Write technical documentation, prepare demos, and present your work in sprint reviews.

Stack & technical environment (reference):

• Language: Python (required, excellent level).
• AI / DL: PyTorch or TensorFlow, scikit-learn, NumPy/Pandas.
• LLM / GenAI:
• Models & tools: Gemini (Vertex AI), open-source (Llama, Mistral…), embeddings, tokenizers.
• Orchestration: LangChain / LlamaIndex / Haystack, LangGraph or agentic frameworks (AutoGen, CrewAI).
• RAG: ingestion (PDF/HTML), chunking, encoding, Vector DB (FAISS, pgvector, Pinecone/Weaviate as needed), reranking (BM25/ColBERT).
• Guardrails & evaluation: Ragas / DeepEval / promptfoo, Pydantic validations, safety rules, basic red-teaming.

Cloud & Dev:

• GCP (preferred): Vertex AI, BigQuery, Cloud Storage, Cloud Run/Functions, Pub/Sub, Secret Manager, IAM.
• Alternatives: AWS (Bedrock, S3, Lambda), Azure (AI Studio, OpenAI, Functions).

MLOps / DevOps:

• Docker, Git/GitHub, CI/CD (GitHub Actions/GitLab CI), testing (pytest), pre-commit, packaging (poetry).

Nice-to-have: basic TypeScript/React for small demo UIs, OpenAPI/REST, security basics (PII, GDPR), LangFuse/Arize Phoenix for tracing.