I'm a husband and father, a researcher at heart, and a computer engineer by profession. I've always been fascinated by technology, and I have a particular love for robotics — hence, Turgibot.
I hold a Master's degree in Computer Science from the Open University of Israel and am currently pursuing a PhD focused on Artificial Intelligence and Robotics. My research interests center around intelligent agents: how they understand their environment, reason about goals and plans, make decisions, and collaborate with humans and other agents.
My path combines academic research with hands-on engineering. Over the years, I've worked as a software engineer and technical lead, conducted and published peer-reviewed research, taught computer science and robotics, and built plenty of systems, prototypes, and experiments along the way.
Turgibot is my personal space on the web — part portfolio, part research notebook, and part knowledge base.
It's where I document what I've learned, built, researched, taught, and achieved. You'll find finished projects and publications alongside technical notes, experiments, ideas, and subjects I'm still exploring.
I don't want the things I learn to disappear into old folders, repositories, papers, or forgotten notes. I want to build a growing, searchable record of my work and knowledge — and, whenever possible, make it useful to others too.
Some things here are finished. Others are still evolving.
That's the point. Turgibot is a record of the journey, not just the results.
Guy Tordjman and Nadav Voloch
DSTRA-GNN predicts ETA with a vehicle-centered dynamic graph and explicit route awareness, cutting error on SUMO and Porto taxi data versus static-graph baselines.
Guy Tordjman and Nadav Voloch
Traffic-DSTG-Gen and TrafficLab generate route-aware dynamic graphs from SUMO and evaluate dynamic GNN ETA models in a live testing environment.
Guy Tordjman and Nadav Voloch
A toolchain that turns SUMO simulations and real trajectories into the same route-aware dynamic graph representation, plus a web tester for ETA models.
Nadav Voloch, Neev Penkar, and Guy Tordjman
A smart real-time estimation method that uses predicted future congestion in simulation to improve ETA over static Dijkstra routing.
Speaking at NiDS was not my first time in front of an audience. It was the first time I presented research that was truly mine — after two bumpy years of work, and with more at stake than the details on the slides.
AI can summarize a research paper in seconds. ReCo was created for something slower and more valuable: helping a reader genuinely understand it.
AI agents changed how I program, and that change forced a harder question — could I keep doing this for the rest of my professional life? This is the story of leaving before I knew where I was going, and finding my way back to research.
A hosted reading workbook for Masters and Sardina's AAMAS 2019 paper on cost-based goal recognition when observed agents are not perfectly rational.
An interactive Block-words simulation of Freedman and Zilberstein's recognizer — likelihood, dynamic prior, necessity conditions, and R_ring's responsive plan — with guided tutorials and the paper's running example step by step.
A hosted reading workbook for Freedman and Zilberstein's AAAI-17 paper, with the PDF kept next to the companion so we can discuss planning, recognition, and responsive interaction together.