Dr Kostas Vasilopoulos

• •Own the engineering strategy and roadmap for Pfizer's enterprise, cloud-native Data Science & GenAI platform (Dataiku), serving 1,800+ data scientists and analysts and hosting a large portfolio of production Python services and web apps
• •Scale LLMOps for 100K+ LLM requests/month, driving cost controls, observability, and access governance across production ML and GenAI workloads
• •Architect enterprise GenAI capabilities (LLM integrations, RAG systems, AI assistants) as reusable platform services that accelerate AI adoption across the organization
• •Establish engineering standards, reusable frameworks, APIs, and self-service tooling that improve developer experience and model lifecycle governance in regulated environments
• •Provide technical leadership across platform, security, data science, and business teams, and manage a direct report

• •Technical owner for integrating and scaling Large Language Models (LLMs) on the enterprise Dataiku platform, solving reliability, performance, and compliance constraints hands-on
• •Designed MLOps execution patterns for model deployment, monitoring, and access control in production, later adopted as platform-wide standards
• •Built reusable platform abstractions and reference implementations for Python services and web apps, driving cross-team adoption of AI capabilities
• •Technical Lead for Ethicara ML: designed and delivered responsible AI solutions aligned with enterprise and regulatory standards
• •Governed platform usage, onboarding, and operations across users, administrators, and engineering teams

Promoted from Data Scientist (06.2021 – 01.2022)

• •Tech Lead for EU Horizon 2020 initiatives (Infinitech, AI4PP), owning end-to-end system architecture, delivery, and long-term technical direction
• •Designed and implemented a distributed data collection and processing platform on Kubernetes, horizontally scaling to millions of jobs per day
• •Productionized multiple ML models with autoscaling, microbatching, and runtime optimizations, increasing throughput by 320% and cutting infrastructure costs by 40%
• •As de facto platform owner, defined reusable ML pipelines and CI/CD, observability, and reliability standards adopted across models, services, and teams
• •Built and operated API microservices integrating PostgreSQL, Elasticsearch, and analytics pipelines for real-time and batch workloads
• •Led development and deployment of production NLP pipelines using Spark NLP (classification, sentiment analysis, NER), later reused across document and news processing workloads

• •Developed and trained NLP models for large-scale news and document analysis, focusing on classification, sentiment analysis, and entity extraction
• •Extended the platform from structured data to unstructured text, laying the groundwork for the production NLP pipelines later scaled as ML Engineer

• •Designed and implemented real-time exuberance detection and predictive modeling systems, translating advanced econometric theory into production-grade open-source software (exuber)
• •Delivered 14–150× performance improvements via optimized matrix inversion and numerical methods, enabling real-time processing of large-scale economic datasets
• •Work from the Housing Observatory featured in The Times and used by practitioners, including central banks

• •Taught undergraduate and postgraduate economics courses, consistently receiving outstanding student evaluations

• •Built analytical models and interactive tools for real-time monitoring of UK national and regional housing markets
• •Led web development of the Housing Observatory platform (https://housing-observatory.com/), collaborating with the Federal Reserve Bank of Dallas
• •Authored and maintained 6 open-source R packages (exuber, ivx, uklr, onsr, transx, ihpdr) with 100K+ collective downloads

Lancaster University

Published research in peer-reviewed journals. Presented work at multiple international conferences and academic workshops.

University of Macedonia

University of Macedonia