CV
Sarbashis Das, PhD
Principal Data Scientist & Technical Lead — Ericsson, Sweden
With over 15 years spanning academic research and enterprise AI, I drive the strategy, architecture, and delivery of AI-powered products at Ericsson. From building an enterprise-grade MCP Gateway and Agentic RAG systems to designing AI agent frameworks adopted across the organization, I operate at the intersection of technical depth and business impact. I set technical direction for cross-functional teams, align AI initiatives with product strategy, and champion the adoption of emerging technologies — including Claude Skills, AI Agents, and Model Context Protocol — to transform how teams build and ship intelligent systems.
Professional Skills
AI & Technical Leadership Technical Vision & Strategy Cross-functional Team Leadership Architecture Decision-Making Stakeholder Alignment Mentorship & Talent Development Build vs Buy Evaluation
Project Management Agile Methodologies JIRA & Confluence Risk Management Resource Allocation Stakeholder Communication
Generative AI & Agents Agentic RAG AI Agents MCP Gateway Claude Skills Prompt Engineering Model Evaluation Agent Orchestration
GenAI Frameworks & Platforms LangChain LangGraph CrewAI AWS Bedrock AgentCore Model Context Protocol RAGAS Haystack
Information Retrieval Semantic Search Vector Search Data Preprocessing Knowledge Graphs Query Understanding
Cloud Platforms Azure ML AWS Bedrock Vertex AI Google Cloud AI Amazon SageMaker
Visualization Kibana Grafana Tableau Power BI
Programming Python R SQL Bash C
Experience
Principal Data Scientist & Technical Lead — AI Agent Platform & Agentic RAG
- Defining and executing the technical vision for Ericsson's AI agent platform, from Agentic RAG architecture to enterprise-grade MCP Gateway with OAuth and multi-server orchestration
- Architected and delivered an MCP Gateway integrating Microsoft Entra ID authentication, enabling secure, scalable tool access across AI agents organization-wide
- Pioneered adoption of Claude Skills and AI agent workflows, establishing reusable patterns for developer productivity across teams
- Presented at AWS re:Invent 2025, showcasing how Ericsson leverages AgentCore to build specialized AI agents that eliminate knowledge silos
- Driving the strategic roadmap for AI-powered products, aligning technical capabilities with business outcomes across multiple stakeholder groups
AI Architect & Technical Lead — Ericsson Product Information Assistant
- Owned the end-to-end architecture and technical direction for a flagship AI product, leading a cross-functional team of 15 across Data Science, Engineering, and Domain expertise
- Shaped the product roadmap in partnership with Product Owners and Strategic Product Managers, translating business needs into technical initiatives
- Delivered 16 successful quarterly releases through disciplined Agile execution, establishing a cadence that became a model for other AI teams
- Drove build-vs-buy decisions and technology evaluations, ensuring the architecture scaled with evolving business requirements
- Established engineering standards, code review practices, and technical mentorship programs that accelerated team velocity
Senior Data Scientist — Global AI Accelerator (GAIA)
- Identified and championed Retrieval Augmented Generation (RAG) as a strategic capability, driving the organization's first production RAG implementation
- Evaluated and benchmarked vector stores, embedding models, and retrieval strategies to establish the reference architecture for RAG at Ericsson
- Designed prompt engineering frameworks and architectural patterns that became the foundation for subsequent Agentic RAG systems
Leaders Core Curriculum Program (LCC)
- Completed Ericsson's 15-week leadership development program focused on executive influence, organizational dynamics, and strategic thinking
- Developed leadership toolkit for driving alignment across distributed teams and influencing without authority
Data Scientist — Global AI Accelerator (GAIA)
- Built end-to-end NLP pipelines for document processing and Knowledge Graph construction, establishing foundational data infrastructure
- Developed custom NER models for Ericsson product domain, improving entity recognition accuracy for downstream search and retrieval
- Designed SPARQL-based query interfaces enabling structured knowledge access across product documentation
Researcher / Senior Researcher
- Led computational biology research applying ML to gene regulation in cardiac disease, publishing in peer-reviewed journals
- Developed unsupervised clustering and dimensionality reduction pipelines identifying 743 dysregulated genes across heart failure subtypes
- Built the analytical foundation and scientific rigor that underpins a data-driven approach to complex systems
Education
- Ph.D. in Computational & System Science
- M.Tech in Computational & System Science
Certifications
- Generative AI and Large Language Model
- Specialization in Machine Learning Engineering for Production (MLOps)
- Specialization in Deep Learning
- Architecting with Google Kubernetes Engine: Workloads
- Architecting with Google Kubernetes Engine: Foundations
- Machine Learning A-Z — Hands-On Python & R in Data Science
- Machine Learning, Stanford University
- Introduction to Data Science in Python, University of Michigan
Awards
- 2024 — Ericsson Key Contributor Award. Recognizes potential to drive Ericsson's long-term strategic priorities.
- 2021 — Ericsson Impact Award. Given for contributions with great impact to Ericsson business value.
- 2017 — Research Grant (co-applicant). Dynamics of gene regulatory networks in response to different stresses in Mycobacteria. Funded by FORMAS.
- 2016 — Research Grant (co-applicant). Non-coding and regulatory RNAs in Mycobacteria. Funded by Swedish Research Council (VR).
- 2012 — Wennergren Postdoc Fellowship. Funded by Wennergren Foundations. Grant offered but gracefully declined.
- 2010 — Senior Research Fellowship. Funded by CSIR, Government of India.
- 2006 — Scholarship for Master of Technology. Funded by DBT, Government of India.
Talks & Presentations
- 2025 — AWS re:Invent, Las Vegas. Showcased how Ericsson leverages AgentCore to build specialized AI agents and domain-focused models.
- 2024 — Ericsson Developer Conference. Presented Ericsson Product Information Assistant with Generative AI.
- 2024 — Uppsala University. Invited talk on Generative AI & LLM to Master Program students.
- 2023 — Multiple Ericsson organizations. Presented & demoed Retrieval Augmented Generation (RAG) Architecture in PIA.
- 2022 — Ericsson Developer Conference, Stockholm. Presented & demoed Ericsson Product Information Assistant.