Spiros Chalkias

AI/ML Engineer

 

Portrait

My journey so far

Jul 2026 — Present

AI Engineer

EPFL — CEDE • Lausanne, Switzerland

Extending the CEDE ecosystem for secure, pedagogical AI tools in education. Designing AI tutors and chatbots powered by locally hosted LLMs, with adaptive tutoring flows, collaborative learning scenarios, and real-time classroom agents integrated with Ed Discussion and course platforms.

EdTech AI Tutors On-Premise AI Adaptive Learning Conversational AI Data Privacy
Jan 2024 — Jun 2026

AI/ML Engineer

Moveo.AI • Athens, Greece

Built production AI agents for enterprise clients across fintech, insurance, and gaming. Focused on performance, cost optimization, and trustworthy agent design, shipping research-backed solutions to production.

Product Agentic Pipeline Proprietary LLMs Context Engineering RAG Voice MCP
Jun 2023 — Feb 2024

Research Fellow

EPFL — NLP Lab • Lausanne, Switzerland

Researched commonsense knowledge representation and reasoning for NLG models under Prof. Antoine Bosselut. Designed an adversarial benchmark to stress-test how well evaluation metrics capture reasoning, and found that all 30+ SoTA metrics — from simple token-overlap ones like BLEU to model-based ones like G-EVAL — reward surface fluency over actual reasoning coherence.

Research NLG LLM Evaluation Adversarial NLP Commonsense Reasoning Benchmarking
Nov 2022 — May 2023

MLOps Engineer

Nokia • Athens, Greece

Built an internal NLP topic modeling tool to automate labeling of system logs and replace manual analysis — designed to help engineers identify patterns and root causes across large volumes of diagnostic data.

Knowledge Discovery System Diagnostics Correlation/Event Analysis
Nov 2021 — Nov 2022

Software Engineer

Nokia • Athens, Greece

Developed 5G Core cloud-native software.

5G Core Cloud-Native Systems Agile
2018 — 2023

B.Sc. in Informatics & Telecommunications

National & Kapodistrian University of Athens

Major: Computer Science
Minor: Computer Engineering & Telecommunications
Specialization: Data & Knowledge Management
Thesis: "An Introduction to Randomizing Deep Polynomial NNs"
Supervisor: Prof. Yannis Panagakis
Voluntary Teaching: "Introduction to Programming" & "Data Structures" courses
Computer Science Telecom Deep Learning Data & Knowledge Management Teaching

Stack

Get in touch

Always happy to connect — drop me a line.