Senior AI Engineer

Bayer Digital Hub Warsaw is developing the Agentic Content Management System (ACMS), a modular, headless, AI-driven content management platform designed for ingesting, validating, structuring, and publishing enterprise knowledge across multiple channels including voice, mobile, chat, conversational interfaces, articles, and knowledge bases.

• Proficiency in Python (FastAPI) or Node.js with strong REST API design.
• Experience with Docker and containerization.
• Knowledge of relational database design (Postgres or equivalent), schema modeling, migrations, and query tuning.
• Experience integrating large language models (Azure OpenAI, OpenAI, Hugging Face, etc.).
• Familiarity with chunking, embeddings, intent extraction, and vector databases.
• Skills in logging/monitoring, structured error handling, unit and integration testing.
• Experience with Git workflows, CI/CD exposure, and awareness of RBAC/security integrations.

Highly Desirable:

• AI Planning & Decisioning: MDP, MCTS usage in agent workflows.
• Big Data: Apache Spark, Hadoop for large-scale processing.
• Streaming & Orchestration: Kafka, Airflow, or equivalents.
• Platform & Scale: Kubernetes for deployment/operations; cloud platforms (Azure, AWS, GCP).
• Frameworks: LangChain, Haystack.
• UI for Demos: Basic React/Next.js for lightweight internal forms.

Microservices & APIs:

• Design, build, and operate production-grade microservices (FastAPI preferred or Node.js) with versioned REST contracts.
• Uphold API-first, headless patterns for multi-channel delivery.

NLP / AI:

• Implement pipelines for text generation, intent/metadata extraction, chunking, and embeddings.
• Integrate vector search as needed (e.g., pgvector, Qdrant, Pinecone).

Data & Storage:

• Model and maintain Postgres schemas (articles, intents).
• Manage migrations, performance, and data quality.

Packaging & Ops:

• Ship Dockerized services with configuration, observability (logging, metrics, tracing), runbooks, and basic SLOs.
• Collaborate with the hub for Kubernetes deployments.

Security & Messaging:

• Implement OAuth2/OIDC and secure service-to-service communication.
• Leverage message brokers where appropriate.

Pipelines & Streaming:

• Contribute to real-time and batch ingestion (e.g., Kafka, Airflow) and/or cloud-native functions (AWS Lambda, Azure Functions, GCP Functions).

Agentic Planning:

• Apply AI planning/decision methods such as Markov Decision Processes (MDP) or Monte Carlo Tree Search (MCTS) to improve reliability and autonomy.

Collaboration:

• Partner with product owner, platform, data, and security teams on requirements, reviews, and releases aligned to enterprise standards.

Specific Project:

• Deliver the Enhanced Article Co-Creation Agent, performing roles of Writer, Intent Structurer, Format Converter, and Self-contained Module for agent consumption and reuse.

• Sport subscription
• Private healthcare
• Canteen
• Free parking
• In-house trainings
• Modern office