Rolan A. ML/Computer Vision
Summary
- More than 10 years’ experience of software development
- Data science skills. Computer Vision, multiple view geometry, camera calibration, LIDAR, object detection, semantic segmentation, instance segmentation, time series, dynamic programming
- Software Engineering skills. Experience of IoT (Internet of Things) and Embedded development
- Solution-oriented scientist focused on R&D and product delivery with 9 years of experience on the outsource domain
- Accustomed to self-education and independent problem solution
- My inspiration is exiting by challengeable and reasonable engineering tasks. Pitching skills from years of conferences attendance and strong understanding of business needs are my strengths
- Intermediate English.
- Availability starting from ASAP
Experience
Technical Lead of Computer Vision
June 2022 – Present
Lead Computer Vision Engineer, Pheon Inc
September 2021 – February 2022
Technical Lead, Intetics Inc
March 2021 – September 2021
Professor Assistant, Karazin KNU
May 2020 – June 2021
Software Engineer, GlobalLogic Inc (Kharkiv, Ukraine)
July 2017 – March 2021
Software Engineer, eCozy Inc
October 2016 – February 2017
Software Engineer, GlobalLogic Inc (Lviv, Ukraine)
September 2015 – October 2016
Researcher, Karazin KNU
January 2013 – May 2020
System Administrator, Ukrnames LLC
January 2012 – January 2013
Projects
Technical Lead, Pet Mask
Jun 2022 – Present
Description: Mobile demo application for cat face and face landmarks detection to apply filter mask. Entertainment video processing system for projecting 3d objects onto pet’s head.
Responsibilities: Leading a team of three CV developers
Technologies: Python, TensorFlow, OpenCV, Face3D
Lead Computer Vision Engineer
Sep 2021 – Feb 2022
Description: Startup in the domain of lip sync and deep fake.
Responsibilities: Supervised and Personalized HQ LipSync R&D. Facial landmark detection. Object detection with R-CNN. Edge computing optimizations (iOS/Android) for real time LipSync. GCP infrastructure architecture and supporting. Time series prediction model with LSTM. TensorFlow vs PyTorch investigation.
Technologies: Python, C++, TensorFlow, Keras, OpenGL, ctype, GCP, PyTorch
Technical Lead
Mar 2021 – Sep 2021
Description: A set of pre-sales in domain of mapping, sensor fusion and 3d reconstruction.
Responsibilities: Object detection with R-CNN/ FasterR-CNN. Image segmentation and object detection. Analytical work with formulas and math models. 3D reconstruction with SLAM from RGBD images. Camera calibration/autocalibration with MVG methods
Technologies: TensorFlow OD API, Keras, Pytorch, Python, CVAT
Professor Assistant, Academic projects
May 2020 – Ongoing
Description: Academic work at a university and science researchers in a lab.
Responsibilities: Analytical work with formulas and math models. Developing of electromagnetic theory.
Technologies: Python, C++, OpenCV, Keras, ctype
Senior Software Engineer
Feb 2019 – Mar 2021
Description: R&D of market-leading automotive vehicle.
Responsibilities: Object detection with TensorFlow OD API. TFv1 to TFv2 migration. Classification of dataset with XGBoost. Analytical work with formulas and math models. Camera calibration/autocalibration with MVG methods. 3D reconstruction with SLAM from RGBD images. Pipeline optimization with Tensor RT and ONNX.
Technologies: Python, C++, NVTX, TensorRT, CVAT, TensorFlow, ctype
Software Engineer
Jul 2017 – Nov 2018
Description: A set of outsource projects in domain of image classification / segmentation.
Responsibilities: PoC prototyping based on CNN image processing. Image segmentation. High load application development with C++.
Technologies: Python, OpenCV, TensorFlow/Keras, C++, TensorFlow Lite, Qt
Software Engineer
Oct 2016 – Feb 2017
Description: Startup in domain of object tracking with edge computing.
Responsibilities: Developing of dashcam with active tracking of vehicles .
Technologies: C++, OpenCV, CMake
Software Engineer
Sep 2015 – Oct 2016
Description: A set of outsource embedded projects in cutting edge domains.
Responsibilities: Development of firmware for IoT network router. Development of smart parking monitoring system.
Technologies: Embedded C, C++, OpenCV, SQL, Bash, CMake
Researcher, Academic projects
Jan 2013 – May 2020
Description: Academic work at a university and science researchers in a lab.
Responsibilities: Analytical work with formulas and math models. Developing of electromagnetic theory. Antenna synthesizing and modeling. Development of software for numerical simulations.
System Administrator, Models rewriting
Jan 2012 – Jan 2013
Description: Web hosting and domain name registry organization.
Responsibilities: System Administration and Client Support.
Technologies: DNS (bind), TCP/IP, KVM, XEN, GNU/Linux, Gentoo, LAMP, CMS (Joomla, Wordpress, Drupal), Python, Bash, RAID, LVM, Iptables, squid, OpenVPN, Jenkins, Gerrit
Pet Projects, and other impact
Significant publication
- Several personal and team patents
- 10+ international conference thesis with SCOPUS impact
- R. Akhmedov, "Neural Radio in DS-UWB IoT Applications", 2020 IEEE Ukrainian Microwave Week (UkrMW), 2020, pp. 1073-1078
- R. Akhmedov, O. Dumin, V. Katrich "Impulse radiation of antenna with circular aperture" Telecommunications and Radio Engineering, vol. 77, pp. 1767-1784, (2018)
UWB-SP Simulation Software
- Transient electromagnetic field simulation
- library-type software with POSIX computability
- High performance parallel distributed computation
- Interruptible computation with controllable progress
- Python interface for C++ software with ctype
- Flexible Cmake build system
- Implementation of EM simulators in time domain: FDTD, nonlinear evolution approach, MMO
- Data exporting as TensorFlow dataset
Smart Surveillance Camera
- Works with Apple HomeKit (no certification)
- Works with most of IP camera aggregators
- FFMPEG multiuser video stream
- Up to 5 parallel video streams
- Motion sensor based on computer vision algorithms
- Configurable motion sensor
- Image depth estimation mode with single RGB camera
Autonomous Surveillance Bot
- SLAM from couple of RGB cameras
- Autocalibration for compass in 3D environment
- 3D vectorized model of environment as output
- Motion and anomaly detection in selected environment
- Automatic finding and wiring with doc station module
- Target haunt and tracking
- Wi-Fi and Cellular wireless interfaces
- TTY-USB debug interface
Education
Postgraduate PhD research program attendance, Universidad de Murcia, Murcia, Spain
2017
PhD in Phys & Math, National Transport University
2010-2020,