Data Scientist

Join our Team

About this opportunity:

Welcome to an exciting opportunity at Ericsson, where you'll be stepping into the role of a Data Scientist. As part of our team, you'll have the opportunity to develop unique machine learning solutions that address complex business issues. If you’ve ever thought “GenAI can do way more in Telecom than chatbots” — this role is for you. We’re building production-grade GenAI systems, AI agents, and enterprise copilots for Telecom OSS/BSS platforms—automation, intelligence, and decisioning at scale. No demos. No fluff. Real systems, real impact.

What you will do:

MCP-based applications and tools integrated with enterprise systems

Building GenAI Agents

RAG pipelines over network, service, customer & operational data

Scalable Python services & APIs powering GenAI workloads

Optimized SLMs for cost-efficient, low-latency deployment

Training of SLM/LLM based on domain requirements.

The skills you bring:

Statistics, Machine Learning Fundamentals, Deep learning expertise.

Feature Engineering & Model Selection.

Strong Python with hands-on ML / NLP and GenAI Agent development E2E experience (production grade not experimentation.

Experience in Training of LLM/SLM models and changing the network Architecture.

Track record of delivering business value in real production grade GenAI Agent.

Solid experience with LLMs, embeddings, retrieval systems and agent orchestration frameworks.

Proven expertise building GenAI copilots, persona-based agents.

Hands-on experience developing applications using MCP.

Experience with production grade cloud-native ML deployments.

Hands on experience in either of AWS Bedrock or GCP Vertex AI.

Good to have:

Strong academic background.

Experience in having journals & patents(US/India).

Telecom OSS/BSS domain ( business and data ) knowledge.

Experience in reasoning , knowledge graph.

Awareness of AI risk, observability, explainability & responsible AI.

Bonus: experience with Databricks, Snowflake, DataRobot, or similar AI platforms.