Tamas Levente Kis

1136 Budapest

Hungary

Mail: tamkis [at] gmail [dot] com

I am a system architect at Ericsson ltd., and the acting CTO of IT-MEDicine Ltd. I was graduated as a medical doctor in 2005 at Medical University of Targu Mures. My main research interests are artificial and biological neuronal networks, and intelligence in general. Myers-Briggs personality type: ENTP-A, but I would argue against labeling me as a debater.

I'm satisfied with my current roles, and not open to a job opportunity unless it is about general AI.

Publications:

	Prediction of general mental ability based on neural oscillation measures of sleep 2005 In Journal of Sleep Research, 2005 Sep;14(3):285-92. My role: quantitative EEG analysis: matlab, dadisp.
	Baroreceptor sensitivity assessed with the finger pulse wave alpha index, light reflection rheography and ambulatory blood pressure monitoring at two minutes 2004 Rom J Intern Med., 2004;42(1):137-42. I was building a hardware, and a software analysing measurements. They've misspelled my name in the publication. They really did.
	A computational model of the visual system Thesis, 2003-2005 A feature-pair based stochastic self-organizing network for object detection and recognition. Translation and light-exposure invariant totally, partially invariant for noise and distortion. I doubt it is available publicly, and neither I have the final version. But I have the code, and I plan to reimplement it in pytorch one day.

Prediction of general mental ability based on neural oscillation measures of sleep

2005

In Journal of Sleep Research, 2005 Sep;14(3):285-92.

My role: quantitative EEG analysis: matlab, dadisp.

Baroreceptor sensitivity assessed with the finger pulse wave alpha index, light reflection rheography and ambulatory blood pressure monitoring at two minutes

2004

Rom J Intern Med., 2004;42(1):137-42.

I was building a hardware, and a software analysing measurements. They've misspelled my name in the publication. They really did.

A computational model of the visual system

Thesis, 2003-2005

A feature-pair based stochastic self-organizing network for object detection and recognition. Translation and light-exposure invariant totally, partially invariant for noise and distortion. I doubt it is available publicly, and neither I have the final version. But I have the code, and I plan to reimplement it in pytorch one day.

Projects:

	zax-json-parser 2020 - PRESENT JSON parser providing a kind of a 'struct to JSON' and 'JSON to struct' conversion. As JSON and C++ structs are totally different constructs, this is a best-effort approach. Though, it could be useful. C++ black magic.
	zax-tensor 2020 - PRESENT Tensor library in C++. It allows direct access to its underlying data buffer, and serializes in JSON. Built on top of zax json parser.
	DataBrowser 2003 - PRESENT Data analysis software easily integratable with runtime analysis systems and hardwares. Not public.
	Feature tracking 2015 Using a simple SOM network for feature tracking. Works suprisingly well.
	Vehicle tracking 2014 Vehicle detection algorithm. No NNs, just simple math. Detects various types of cars regardless of their color, robust against fluctuating light intensity and shadows.
	License plate detection 2003 Algorithm detecting license plates. Just a banch of spatial FIR filters with different kernel shapes, not necessarily rectangulars.
	Student years Image montage of my student work.