CV
Last updated May 2026.
Contact Information
| Name | Charidimos (Harry) Papadakis |
| Professional Title | Incoming PhD Student, MIT Operations Research Center (Fall 2026) |
| harrypapadakis02@gmail.com | |
| Phone | +30 6976139094 |
Education
-
2026 - 2031 (expected) Cambridge, MA, USA
PhD
Massachusetts Institute of Technology
Operations Research (MIT ORC)
- Advisor: Prof. Dimitris Bertsimas
-
2020 - 2025 Athens, Greece
MEng (Diploma, 300 ECTS)
National Technical University of Athens
Electrical & Computer Engineering
- Diploma Thesis: “Adaptive Multi-Agent LLM Systems for Financial Trading: A Framework for Realistic Simulation and Dynamic Prompt Optimization”
- Supervisor: Prof. Giorgos Stamou
Experience
-
May 2025 - Present Athens, Greece
AI Software Engineer
Veltiston AI
- Designed an agentic RAG system for document analysis, deployed as a Microsoft Word add-in; coordinated multi-step retrieval over external document corpora with grounded citation generation.
- Developed agent orchestration for query decomposition, evidence aggregation, and answer synthesis under context-length and latency constraints.
- Contributed to a hospital operating-room scheduling platform: defined KPIs and built data pipelines from clinical systems to support downstream optimization.
-
Jul 2024 - Sep 2024 Athens, Greece
Software Engineering Intern
Netcompany–Intrasoft
- Contributed to backend services for the Austrian Customs Clearance System, a large-scale public-sector software project.
-
Jul 2023 - Sep 2023 Athens, Greece
AI Engineering Intern
RiverTech
- Developed machine learning models on IoT smart-home data for behavioral pattern recognition, supporting energy optimization research.
Publications
-
2026 ATLAS: Adaptive Trading with LLM Agents Through Dynamic Prompt Optimization and Multi-Agent Coordination
ACL 2026 (Main Conference)
An adaptive prompt-optimization framework that improves coordination and decision quality among LLM-based trading agents in dynamic markets.
-
2025 StockSim: A Dual-Mode Order-Level Simulator for Evaluating Multi-Agent LLMs in Financial Markets
Preprint (planned EMNLP 2026 System Demonstrations)
A high-fidelity, dual-mode simulation environment for reproducible evaluation of LLM agent strategies and behavior under realistic order flows.
Skills
Programming: Python, C/C++, Java, JavaScript/TypeScript, SQL
LLMs & Agentic AI: LangChain, LangGraph, Hugging Face Transformers, Prompt Engineering & Optimization, RAG, Vector Databases, Multi-Agent Orchestration, MCP
ML / Deep Learning: PyTorch, TensorFlow, Scikit-learn, Ray, NumPy, pandas, matplotlib
Backend & Web: FastAPI, Spring Boot, Node.js, Next.js, React, REST APIs, Celery
Systems & Tooling: Docker, RabbitMQ, MySQL, AWS S3, Git, Linux/Unix
Languages
Greek : Native
English : C2
German : C1