top of page
Soyut Arkaplan


Co-founder & CTO at Ravinspect





  • LinkedIn
  • Ravinspect
  • Ravinspect
  • Scholar

Istanbul / Turkey


English (C1), Turkish (Native)

**You can also view my online resume at

Co-founder of Ravinspect and working as a CTO since 2018. Ravinspect develops autonomous indoor UAV systems for warehouses that enables unmanned stock control operations. Responsible for R&D operations related to autonomy software, computer vision, and system design. Ravinspect was accomplished to get multiple government funds and the product find a place in the market. The system is operational in Turkey and USA at one of the largest home appliances companies.

Along with entrepreneurship, academic career is ongoing as a Ph.D. candidate, and published papers can be found in the related section.




Co-founder & CTO


Developed indoor navigation mechanism, pallet label detection & recognition models, GUI for drone control as a ground station, and autonomous flight, path planning, collision avoidance, failsafe detection, trajectory manipulation algorithms. Led the computer vision and robotics teams



VLSI Laboratory / Istanbul Tech. Uni.

Developed a cryogenic modeling methodology for MOSFETs to use in space applications. This project is funded by the government.






Istanbul Technical University

Thesis: Predictive anomaly detection on UAVs



Master's Degree

Istanbul Technical University

Thesis: Cryogenic modeling methodology on MOSFETs


Bachelor's Degree


Istanbul Technical University



C++ (5+ years) / Python (4+ years)

Linux (5+ years daily basis)

OpenCV / Keras / PyTorch

4+ years Computer Vision experience

ROS1 (5+ years)

Git (5+ years) / Docker (1+ year)

PyQT (3+ years)

Experienced Hardwares:

Jetson TX2/Xavier NX, Realsense T265/D435i, ZED1/2, MIPI Cameras, Pixhawk(PX4), LIDARs



All these projects were contributed as both developer and the manager.

You can reach more details of each project by clicking the titles


Used VIO with LIDAR assistance and EKF. The position error is below %0.2. Designed specifically for warehouses. Apriltags are used for trajectory correction at each 50m.

Screenshot from 2020-12-30 11-37-52.png

Ground control station to create stock counting missions by the selected counting area. Users can visualize the instant location of the drone and the whole warehouse. The detected labels can be inspected and the stocktaking reports can be created automatically.


PX4 autopilot used with companion computer. No need for external communication or RC. Fly indoors without GPS. Redundant design with added extra failsafe checks and collision avoidance mechanisms.


Pallets may have multiple labels. Object detection models can distinguish the desired label and text detection/recognition models find barcodes. All 3 models run on the drone. A lot of optimization was implemented to run with localization and autonomous flight algorithms.


Depth camera and LIDAR based mapping. Rack identification and trajectory correction using last-seen data. Online mapping/planning. No need of predefined maps.

Screenshot from 2021-12-22 22-08-20_edited.jpg

Precise landing feature. Passive electrical contacts that can be activated when the drone is landed. It is triggered autonomously by a drone. No need for any human intervention. Certified safety.



  • Kabaoglu, A., Solmaz, N. S., Ilik, S., Uzun, Y., & Yelten, M. B. (2018). Statistical MOSFET modeling methodology for cryogenic conditions. IEEE Transactions on Electron Devices, 66(1), 66-72.

  • Kabaoglu, A., & Yelten, M. B. (2017, June). A cryogenic modeling methodology of MOSFET IV characteristics in BSIM3. In 2017 14th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD) (pp.1-4). IEEE.

  • Kabaoglu, A., Sahin-Solmaz, N., Ilik, S., Uzun, Y., & Yelten, M. B. (2019). Variability-aware cryogenic models of mosfets: validation and circuit design. Semiconductor science and technology, 34(11), 115004.

  • Ilik, S., Kabaoglu, A., Solmaz, N. S., & Yelten, M. B. (2019). Modeling of total ionizing dose degradation on 180-nm n-MOSFETs using BSIM3. IEEE Transactions on Electron Devices, 66(11), 4617-4622.

  • Uzun, Y., Kabaoglu, A., & Yelten, M. B. (2019, November). Design of a LC Voltage-Controlled Oscillator for Space Applications in C-Band. In 2019 11th International Conference on Electrical and Electronics Engineering (ELECO) (pp. 388-391). IEEE.

  • Kayıhan, H. I., Kabaoglu, A., & Yelten, M. B. (2019, November). A Cryogenic CMOS Low Dropout Voltage Regulator Design for Space Applications. In 2019 11th International Conference on Electrical and Electronics Engineering (ELECO) (pp. 392-396). IEEE.

  • Ilık, S., Solmaz, N. S., Kabaoglu, A., & Yelten, M. B. (2019, July). Comparison of ELTs with different shapes and a regular layout transistor in 180 nm CMOS process. In 2019 16th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD) (pp. 21-24). IEEE.



After graduation, I worked as a research student at the VLSI laboratory and participated in the "Cryogenic Electronic Applications in Space" project, developing software models that resulted in several publications. Upon completing the project, I began my entrepreneurship journey as a co-founder of Ravinspect. 

Our first project aimed to inspect aircraft for lightning strike damage. To achieve this, we developed a precise indoor navigation system and AI models to detect the damage. The precision of the indoor navigation system was outstanding for a drone at that period and it was one of the pioneers in the field. We conducted demos with Turkish Airlines and Pegasus using our prototypes, but eventually shifted to another industry due to aviation regulations.


We maintained the core of the drone while modifying our computer vision approach to detect labels, barcodes, and text. We made the first sale to one of the largest home appliances companies B/S/H in a short period of time after the adaptation.

During this time, we received three government funds from the Scientific and Technological Research Council of Turkey and received support from the best incubation center in Turkey. Ravinspect was recognized as a deep tech startup in Turkey by HelloTomorrow and joined the TekeOff Startup Summit, NVIDIA Inception, TIM-TEB Startup House, 5G@EndTech programs. We have also exhibited our product at the CES expo. Additionally, Ravinspect has become an exporter by entering the USA market.

bottom of page