About the opportunity
Ericsson’s Service Architecture team, within Service Management in the Operate Service Line, is building intelligent automation and decision-support systems for telecom-scale managed services. In Software Development, you will design, build, test, and ship production-grade AI services focused on Agentic AI, LLM integrations, vector retrieval, knowledge graph reasoning, and ML infrastructure. Working with platform engineers, data scientists, and product owners, you will turn AI experiments into reliable services that deliver measurable value across Ericsson’s managed service operations.
What you will do
-
Design multi-agent systems using LangGraph, AutoGen, CrewAI, Google ADK, or equivalent orchestration.
-
Build LLM services with prompt engineering, tool/function calling, planning, memory, reflection, guardrails, fallback patterns, and RAG.
-
Integrate and evaluate commercial API-based and open-source/self-hosted LLMs, balancing cost, latency, accuracy, privacy, and complexity.
-
Build RAG pipelines using ChromaDB, Qdrant, Vertex AI RAG, Pinecone, Weaviate, pgvector, Milvus, or equivalent; manage embeddings, hybrid search, re-ranking, and index lifecycle.
-
Model and query property graphs and RDF/ontologies to enrich agent context, semantic reasoning, and structured retrieval.
-
Create evaluation harnesses with automated scoring, human review, A/B testing, and golden datasets.
-
Develop ML pipelines for feature engineering, training, evaluation, serving, fine-tuning, packaging, versioning, and monitoring using Python, PyTorch, scikit-learn, and HuggingFace.
-
Apply ML methods including regression, classification, tree ensembles, boosting, SVMs, neural networks, clustering, PCA/UMAP, LSTMs, and Transformers.
-
Monitor model drift, data quality, performance, latency, and token cost in production.
-
Build clean, testable Python/PySpark services and REST/FastAPI/Flask/gRPC APIs using clean-architecture principles.
-
Contribute to CI/CD with GitHub Actions, Jenkins, or ArgoCD; participate in reviews, architecture discussions, documentation, runbooks, and mentoring.
You will bring
-
6 –15 years of software development experience and a Bachelor’s/Master’s degree in Computer Science, Software Engineering, AI/ML, or equivalent.
-
5+ years of hands-on ML engineering and LLM development experience in production.
-
Production experience with Agentic AI, vector database-backed RAG, and knowledge graph integrations.
-
Expert Python; PySpark is desirable. Experience with Spark, SQL, Pandas/Polars, structured streaming, Delta Lake/Iceberg, and Kafka/Pulsar.
-
LLM fine-tuning and optimisation knowledge, including LoRA/QLoRA, RLHF, ONNX, TensorRT, quantisation, and efficient inference.
-
Experience with Docker, Kubernetes, AWS/Azure/GCP, Databricks or EMR, and MLOps tools such as MLflow or Weights & Biases.
-
Knowledge of LLM observability, structured logging, token-cost tracking, alerting, responsible AI, and model safety.
-
Working knowledge of TOGAF or an equivalent enterprise architecture framework and technical documentation standards.
